Just my two cents worth on the release being held back by Windows;
I think that, regardless of your position on Windows as an operating system, or Windows users, that putting out a release that is crippled would be downright humiliating to the *whole* Inkscape project.
Let's ask, what is pushing us so hard to get the next release out? In a lot of these projects there is an impetus based on saving face, which is about bringing new features to fruition. IMHO, the saving face needs to be done by being a little patient, or at the very least not releasing 0.46 for Windows until it is pretty right.
Again, what is the aim? Is it purely to further the name of Inkscape, or is there at least some desire to further SVG as a viable option is a largely proprietary world? I have heard that said a number of times, and used as a reason to include or reject features. This project is meant to be about a good SVG editor, and the rest is a nice by-product. Again, a good reason to hold off just a little.
Let us also ask, if Linux users are so technically superior, aren't they already using the release candidate, which offers pretty much what the release will? Surely even the nightlies are very stable by now if all the bugs are ironed out. If so, in at least one sense who cares if there's a release if the RC is working and those who want the release are already using it! Linux users are well and truly used to beta and release candidates, and in fact the pure only use the bleeding edge stuff and look down on those who don't. Or are we next going to start railing against Ubuntu users for their lack of technical expertise, and because they need a packaged version? Afterall, isn't Ubuntu trying to attract that Windows-like person?
Lets please hold off just a little, at the very least on the Windows release. For many Linux users it makes no difference whether or not there's an official release, because the pre2 package is stable. For the Windows users, and the reputation of Inkscape, SVG, and even FOSS, it makes a huge difference.
Remember too, that we are mainly talking about one bug, the https://bugs.launchpad.net/inkscape/+bug/179988 ability to print , which IMHO must be working for a Windows release. A lot of the other things like native dialogues (which is close) or OCAL importing (mabe not so close) can be reverted or turned off if they are causing problems, at least until the 0.46.1 release (which may even be Windows only). Even upside down text in PDF export can be fudged by forcing text to path until that is fixed. But no printing is pretty much life or death for most people.
Native dialogs will not be shipped with 0.46. I committed them only after the 0.46 build was branched.
On Sat, 2008-03-01 at 03:05 -0800, rygle wrote:
Remember too, that we are mainly talking about one bug, the https://bugs.launchpad.net/inkscape/+bug/179988 ability to print , which IMHO must be working for a Windows release. A lot of the other things like native dialogues (which is close) or OCAL importing (mabe not so close) can be reverted or turned off if they are causing problems, at least until the 0.46.1 release (which may even be Windows only). Even upside down text in PDF export can be fudged by forcing text to path until that is fixed. But no printing is pretty much life or death for most people.
Thanks Aaron,
I agree with lots of your comments, but I think you have somewhat misunderstood me. I too am very thankful for FOSS software, and the many people behind it. I am merely pleading that we don't release a *Windows version* that is crippled, particularly in the area of printing (see the bug I mentioned).
All I am asking is that there be patience with the *windows* release, because for those who aren't as technically astute, a release that can't print is going to scare them off. Hence the comment about humiliating. And I think not just Inkscape, but FOSS, and SVG too.
Aaron Spike wrote:
A Windows-less release is only humiliating to the Windows using portion of the community.
Not what I was saying. A windows-crippled release, which has been strongly suggested in the area of printing, would scare a lot of people away. I just gave out fifteen CDs with Inkscape 0.45.1 this week, and I didn't give out the Dev version, because I knew it wasn't ready. I would not give out a print crippled version of 0.46, because there's a lot of negative PR in that sort of thing. That's what I meant about humiliation.
Please explain how you have arrived at the supposition that the motivation for this Inkscape release is about "saving face"? It is a goal of our project to have timely releases. This is a good goal for every OSS project. It helps projects to move forward and gauge progress. And it helps us slow down and focus on identifying and resolving a large number of issues with functionality and stability.
I agree and support moving forward. I guess I feel very discouraged that members of the community have started to use windows people as the whipping boy that is holding back the release. I would even describe it as railing against them. Windows people are still people. Non-programmers are still people. And when people are strongly suggesting that we should just drop printing, after several other rants, it makes me really start to question motives.
I find this paragraph insultingly stereotypical. ...I believe that the arguments you have made, because of their tone, do more to distract from your argument rather than support it.
Well, I feel like my tone is a lot gentler than several comments I've seen lately from others, and felt I needed to ask for some patience. I don't demand anything from any developers, and I am not asking anyone to spend money or time, but I *ask* for some patience from some on the windows version.
I probably didn't say clearly enough that I don't mind a late release on Windows, just not a crippled release in such an important area as printing. Hopefully that is now clear, and hopefully it would mean that no-one would have to rant, me included. </flame>
Speaking of releases, I still really want to see tref working (or if not, then the release notes need to be updated). I have been suddenly swamped with school work so my compile fun to find the commit that caused the issue has been temporarily halted.
So, my question is, what are we looking at for a release date at this moment, and does anyone want to help me at least narrow down the commit? :) Right now I believe everything worked after Rev 15430, and doesn't work by 15362. If you are willing to help me, please email off list and you will be forever my best friend!! I should be able to then look at the problem a lot more easily if I can narrow down the commit a bit more.
Gail
On Sat, Mar 01, 2008 at 10:34:30AM -0500, Gail Carmichael wrote:
So, my question is, what are we looking at for a release date at this moment?
Fedora 9 goes into feature freeze and beta freeze March 4th, and beta release March 13th with Final Development freeze April 8.
Ubuntu Hardy entered feature freeze Feb 14, and goes to Beta Freeze March 13th with beta release March 20th.
For inclusion for both Fedora and Ubuntu, we'd need to release by the 4th. If we skipped Fedora, we'd have until the 13th. If we miss both of those dates, then the next opportunity to get Inkscape out officially to Linux users would be 6 months from now.
Bryce
I would promote the latest version.
So I vote for release 0.46 without windows and shift a 0.46.1 ASAP after.
Adib. ---
Bryce Harrington schrieb:
On Sat, Mar 01, 2008 at 10:34:30AM -0500, Gail Carmichael wrote:
So, my question is, what are we looking at for a release date at this moment?
Fedora 9 goes into feature freeze and beta freeze March 4th, and beta release March 13th with Final Development freeze April 8.
Ubuntu Hardy entered feature freeze Feb 14, and goes to Beta Freeze March 13th with beta release March 20th.
For inclusion for both Fedora and Ubuntu, we'd need to release by the 4th. If we skipped Fedora, we'd have until the 13th. If we miss both of those dates, then the next opportunity to get Inkscape out officially to Linux users would be 6 months from now.
Bryce
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
At 12:26 PM 3/2/2008 +0100, theAdib wrote:
I would promote the latest version.
So I vote for release 0.46 without windows and shift a 0.46.1 ASAP after.
Adib.
I'm not a developer, but I tend to agree. If you'll pardon a mangled metaphor, let's get on with it and let the chips fall where they may.
- Alvin Penner
On Sun, Mar 2, 2008 at 7:26 AM, theAdib <theadib@...1439...> wrote:
I would promote the latest version.
So I vote for release 0.46 without windows and shift a 0.46.1 ASAP after.
I disagree. This will defuse the bang. Piecemeal one-platform-only releases will fizzle without attracting much attention. Remember, it's not an availability issue: those who need it can get it any time already. Making a release is primarily a way to make noise and attract new users (and eventually developers). For this, we must run a single concentrated, polished, coordinated campaign. This means no missing platforms and no severe regressions, among other things.
[lurk mode=off] Can't you sneek it into the Linux distros with nearby deadlines without making a big splashy announcement? --bb
On Mon, Mar 3, 2008 at 6:36 AM, bulia byak <buliabyak@...400...> wrote:
On Sun, Mar 2, 2008 at 7:26 AM, theAdib <theadib@...1439...> wrote:
I would promote the latest version.
So I vote for release 0.46 without windows and shift a 0.46.1 ASAP after.
I disagree. This will defuse the bang. Piecemeal one-platform-only releases will fizzle without attracting much attention. Remember, it's not an availability issue: those who need it can get it any time already. Making a release is primarily a way to make noise and attract new users (and eventually developers). For this, we must run a single concentrated, polished, coordinated campaign. This means no missing platforms and no severe regressions, among other things.
-- bulia byak Inkscape. Draw Freely. http://www.inkscape.org
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On 3/2/08, bulia byak <buliabyak@...400...> wrote:
On Sun, Mar 2, 2008 at 7:26 AM, theAdib <theadib@...1439...> wrote:
I would promote the latest version.
So I vote for release 0.46 without windows and shift a 0.46.1 ASAP
after.
I disagree. This will defuse the bang. Piecemeal one-platform-only releases will fizzle without attracting much attention. Remember, it's not an availability issue: those who need it can get it any time already. Making a release is primarily a way to make noise and attract new users (and eventually developers). For this, we must run a single concentrated, polished, coordinated campaign. This means no missing platforms and no severe regressions, among other things.
First, I'm not suggesting I deserve a vote in this decision, I just want to offer up my point of view FWIW. Bulia is right about releases being marketing events, especially for companies that build their brand around that sort of thing like VisualStudio, but I doubt that Inkscape can attract more graphic design users through a release than getting the app into the hands of two major Linux distros. Six months is a serious chunk of market penetration time in web years. If you are going after developers you are doing a good job with SVN and your wiki. As a new user, I'm going to jump on based on my needs and who well I see them expressed in Inkscape. I think you could do a lot more on your front page to blow your horn. I've been pleasantly surprised to find just how powerful a platform you've built upon and into and around called Inkscape. Your better application architects than you give yourselves credit for. Respectfully, Mike Yinger, http://www.ticketprinting.com
On 3/2/08, Michael Yinger <myinger@...1834...> wrote:
On 3/2/08, bulia byak <buliabyak@...400...> wrote:
On Sun, Mar 2, 2008 at 7:26 AM, theAdib <theadib@...1439...> wrote:
I would promote the latest version.
So I vote for release 0.46 without windows and shift a 0.46.1 ASAP
after.
I disagree. This will defuse the bang. Piecemeal one-platform-only releases will fizzle without attracting much attention. Remember, it's not an availability issue: those who need it can get it any time already. Making a release is primarily a way to make noise and attract new users (and eventually developers). For this, we must run a single concentrated, polished, coordinated campaign. This means no missing platforms and no severe regressions, among other things.
First, I'm not suggesting I deserve a vote in this decision, I just want to offer up my point of view FWIW. Bulia is right about releases being marketing events, especially for companies that build their brand around that sort of thing like VisualStudio, but I doubt that Inkscape can attract more graphic design users through a release than getting the app into the hands of two major Linux distros. Six months is a serious chunk of market penetration time in web years. If you are going after developers you are doing a good job with SVN and your wiki. As a new user, I'm going to jump on based on my needs and who well I see them expressed in Inkscape. I think you could do a lot more on your front page to blow your horn. I've been pleasantly surprised to find just how powerful a platform you've built upon and into and around called Inkscape. Your better application architects than you give yourselves credit for. Respectfully, Mike Yinger, http://www.ticketprinting.com
BTW, I know the difference between "your" and "you're", I just can't teach my fingers.
Here is how we can solve this for the release.
Bulia, the new standard Windows dialog was implemented by Kees in Nov/Dec/Jan in r16655, r16860, r16865, r16866 (and maybe a couple others I missed.) Please test reverting these and prepare a patch-set which removes the new print system and restores the old one. Post this patch to bug 179988 and assign it to Ishmal. Please get this done and tested as soon as possible (today or tomorrow preferrably). Hopefully it should not take very long.
Kees, please let Bulia know of anything special that may help him with making this patch, or if you have ideas for a better workaround. If you could clarify which patches need reverted, that could help immensely.
Ishmal, please include this patch in the Windows package. I don't know if you have a deb/rpm-style patching system, but if not, just go ahead and patch source. From your research, it sounds like this will be fixed with the cairo 1.6 release; I asked cworth about a release date for this, and he said it should be within a few weeks, perhaps less, so we probably won't need to carry this reversion patch for very long.
If we can get this reversion patch by tomorrow, then we can avoid dealing with the question of whether to delay, release without windows, etc. etc. It's good to see all the discussion around this (and I'll post my own thoughts later), but at this point what we need first is a patch.
Bryce
P.S., while for Windows it may not an availability issue, it *IS* going to be an availability issue for Linux if we miss the Fedora 9 and Ubuntu Hardy deadlines. Shipping rc's or 0.45.1 in those, when we're so close now, would be embarrassing.
On Sun, Mar 2, 2008 at 7:31 PM, Bryce Harrington <bryce@...1798...> wrote:
Here is how we can solve this for the release.
Bulia, the new standard Windows dialog was implemented by Kees in Nov/Dec/Jan in r16655, r16860, r16865, r16866 (and maybe a couple others I missed.) Please test reverting these and prepare a patch-set which removes the new print system and restores the old one.
I don't think I will be able to test it properly - I don't have a printer hooked up to my Windows laptop so I cannot really try printing from Windows.
Besides, I'm not sure this is the best approach. I think we should better spend the resources on testing the gtk/cairo patch and, if it works, building windows lib package with it.
And in general, I don't really feel like we're ready to release 0.46 tomorrow on ANY platform. Too little time in the freeze. Too many crash bugs still being reported. Just one thing, Ted's extension fixes are not ready yet (issues remain, what was fixed is not yet tested well enough) and not even in the 0.46 branch. Even the release notes are not finalized yet!
Sorry but this is how I perceive it. If these distro deadlines are so important (I never heard about them until today), someone should have made huge noises about them much earlier. (If the noises were being made and I just missed them, sorry, but from all the release process so far I didn't get the idea of urgency, it seemed quite laid-back to me.) So, painful as it is, I just don't think we are ready.
This is a lesson for all of us. We pride ourselves on the "always working" SVN, but in this case, the printing was known not to work, and was not even looked into, for many months. Perhaps everyone was just thinking, "we'll fix it for the release for sure, for now I want to play with something less boring" - and now the release is tomorrow and we're forced to make ugly rushed workarounds. Let's not allow such things happen again.
bulia byak wrote:
On Sun, Mar 2, 2008 at 7:31 PM, Bryce Harrington <bryce@...1798...> wrote:
Here is how we can solve this for the release.
Bulia, the new standard Windows dialog was implemented by Kees in Nov/Dec/Jan in r16655, r16860, r16865, r16866 (and maybe a couple others I missed.) Please test reverting these and prepare a patch-set which removes the new print system and restores the old one.
I don't think I will be able to test it properly - I don't have a printer hooked up to my Windows laptop so I cannot really try printing from Windows.
You can print to a file using a postscript driver and view the output in ghostview. There are a number of packages that will automate this for you and create a too, such as PrimoPDF. This is how I've been testing it because I don't have a printer at home.
Aaron Spike
On Sun, Mar 02, 2008 at 09:01:36PM -0400, bulia byak wrote:
On Sun, Mar 2, 2008 at 7:31 PM, Bryce Harrington <bryce@...1798...> wrote:
Here is how we can solve this for the release.
Bulia, the new standard Windows dialog was implemented by Kees in Nov/Dec/Jan in r16655, r16860, r16865, r16866 (and maybe a couple others I missed.) Please test reverting these and prepare a patch-set which removes the new print system and restores the old one. Post this patch to bug 179988 and assign it to Ishmal. Please get this done and tested as soon as possible (today or tomorrow preferrably). Hopefully it should not take very long.
If we can get this reversion patch by tomorrow, then we can avoid dealing with the question of whether to delay, release without windows, etc. etc.
I don't think I will be able to test it properly - I don't have a printer hooked up to my Windows laptop so I cannot really try printing from Windows.
Since this task has not been achieved, we seem to have little choice than to take the alternative of releasing 0.46.0 as a Linux/OSX only release, with official Windows support being postponed to 0.46.1. Several people including Bob Jamison have spoken up for this alternative, and I hope the silver lining of this cloud will be stimulating more windows folk to get involved in the maintenance of the windows port.
Thanks, Bryce
On Sun, 2 Mar 2008 21:01:36 -0400, "bulia byak" <buliabyak@...400...> wrote:
This is a lesson for all of us. We pride ourselves on the "always working" SVN, but in this case, the printing was known not to work, and was not even looked into, for many months. Perhaps everyone was just thinking, "we'll fix it for the release for sure, for now I want to play with something less boring" - and now the release is tomorrow and we're forced to make ugly rushed workarounds. Let's not allow such things happen again.
I think this is one area where the longer release cycles have hurt us recently -- it's been too easy to become complacent about things like that without frequent releases to keep us honest.
-mental
On Mon, Mar 03, 2008 at 02:23:44PM -0800, MenTaLguY wrote:
I think this is one area where the longer release cycles have hurt us recently -- it's been too easy to become complacent about things like that without frequent releases to keep us honest.
Agreed. The long periods between releases are in part due to how long it takes to get through releases. I started the planning for the release back in November, and here we're on month 4 now. For me the best thing about the release will be finally getting my freetime back!
It's also quite stressful to have to say no to good new features, and to worry over a pile of bugs that doesn't shrink fat enough. Thankfully the new bug tracker makes life a lot easier; with SourceForge, by this stage in the release I'd be completely burnt out.
We really ought to be releasing every 6 months. Every 3 months would probably be even better. This would mean that we'd need to be able to go from Slush to Release in under 4 weeks.
To achieve that, we'd need to have faster response on bugs, scale feature work down (or make better use of branches), and use test suites and automation more heavily. As Ishmal pointed out, we also need more maintainers for caring for platform bugs. And since a large proportion of our release critical bugs were in relation to extensions, we need a stronger team and a better approach there.
I also think if several people were to focus exclusively on bug fixes instead of feature work for a release or two, and get our bug count down to something reasonable, it would make releases a breeze to get through.
Bryce
Bryce:
Wow, I wish that we could harness all of this email-writing energy on the topic, and focus it on the actual work.
You were absolutely right on your blog. The ratio of developers to users on win32 is much lower than on Linux. If we had just one or two more developers, it would -double- our abilities.
Since the win32 stuff has never been a release in its own right, but just an alternate way of building Inkscape, I don't see any problem with releasing Inkscape immediately, and getting a win32 build out whenever it's ready. People can always get a copy off the server. I mean, we build it several times a day. Just pick a build and package it. And if the users don't like it, then we can refund their money. I am a big fan of Fedora, and would -love- to see 0.46 in Fedora 9.
I'll see what patches we need to get printing working well again soon. Maybe we can do the switch to using the new Cairo1.5+ win32_printing_surface in a patch ourselves, rather than wait for Gtk to do it. If so, then maybe a Cairo guy could help us. There ya go. A good opportunity to delegate. :-)
I have been slightly annoyed recently with the vaguely implied idea that we don't -want- to fix this bug, and that any problems we have with Inkscape must be a symptom of our laziness and disrespect for the user. I can assure people that considerable effort has already been spent on this. I have been hoping that the increased clamoring would be accompanied with increased volunteering. I haven't seen it yet, but Hope Springs Eternal.
bob (ishmal)
Bryce Harrington wrote:
Here is how we can solve this for the release.
Bulia, the new standard Windows dialog was implemented by Kees in Nov/Dec/Jan in r16655, r16860, r16865, r16866 (and maybe a couple others I missed.) Please test reverting these and prepare a patch-set which removes the new print system and restores the old one. Post this patch to bug 179988 and assign it to Ishmal. Please get this done and tested as soon as possible (today or tomorrow preferrably). Hopefully it should not take very long.
Kees, please let Bulia know of anything special that may help him with making this patch, or if you have ideas for a better workaround. If you could clarify which patches need reverted, that could help immensely.
Ishmal, please include this patch in the Windows package. I don't know if you have a deb/rpm-style patching system, but if not, just go ahead and patch source. From your research, it sounds like this will be fixed with the cairo 1.6 release; I asked cworth about a release date for this, and he said it should be within a few weeks, perhaps less, so we probably won't need to carry this reversion patch for very long.
If we can get this reversion patch by tomorrow, then we can avoid dealing with the question of whether to delay, release without windows, etc. etc. It's good to see all the discussion around this (and I'll post my own thoughts later), but at this point what we need first is a patch.
Bryce
P.S., while for Windows it may not an availability issue, it *IS* going to be an availability issue for Linux if we miss the Fedora 9 and Ubuntu Hardy deadlines. Shipping rc's or 0.45.1 in those, when we're so close now, would be embarrassing.
On Sun, Mar 2, 2008 at 9:59 PM, Bob Jamison <rwjj@...127...> wrote:
I'll see what patches we need to get printing working well again soon. Maybe we can do the switch to using the new Cairo1.5+ win32_printing_surface in a patch ourselves, rather than wait for Gtk to do it. If so, then maybe a Cairo guy could help us. There ya go. A good opportunity to delegate. :-)
So, can you do a quick build of libs package with this patch?
http://bugzilla.gnome.org/attachment.cgi?id=99752&action=view
If you do, I will do whatever testing I can on Windows today/tomorrow.
I have been slightly annoyed recently with the vaguely implied idea that we don't -want- to fix this bug, and that any problems we have with Inkscape must be a symptom of our laziness and disrespect for the user. I can assure people that considerable effort has already been spent on this.
Of course many people deserve highest possible kudos for their hard work. In particular I don't know what we would be doing without your lib packages. Still, we have what we have with this bug, and this should be a wake up call.
I have been hoping that the increased clamoring would be accompanied with increased volunteering. I haven't seen it yet, but Hope Springs Eternal.
You're not right. Just compare the frequency of SVN commits now versus just a year ago. The growth in development is palpable. However, this is one reason that the problem of bugs and regressions is so urgent. We really must do something to address this, such as implement automated testing of at least some aspects. Now that we have all verbs available via command line, I rather wonder why it's not yet done :)
On Mar 2, 2008, at 6:26 PM, bulia byak wrote:
You're not right. Just compare the frequency of SVN commits now versus just a year ago. The growth in development is palpable. However, this is one reason that the problem of bugs and regressions is so urgent. We really must do something to address this, such as implement automated testing of at least some aspects. Now that we have all verbs available via command line, I rather wonder why it's not yet done :)
Ummm...
a big, huge "ditto!"
bulia byak wrote:
You're not right. Just compare the frequency of SVN commits now versus just a year ago. The growth in development is palpable. However, this is one reason that the problem of bugs and regressions is so urgent.
But what this upsurge is all about, is people working on their own projects. What I'm talking about is supporting Inkscape in general.
For example, my areas of contribution -should- be DOM, binding, and tracing. Yet I never seem to have any time to work on these, with the daily calls for Win32 stuff. I very much want to cut my time on w32 by at least 75%.
We -really- need volunteers to help with the mundane, boring, thankless stuff like maintaining up-to-date dependencies, finding bugs, doing builds, and being a liaison with Gtk and Cairo. We really need to broaden the bandwidth on this, to avoid the inevitable contributor crash and burnout.
bob
On Sun, Mar 2, 2008 at 11:18 PM, Bob Jamison <rwjj@...127...> wrote:
But what this upsurge is all about, is people working on their own projects. What I'm talking about is supporting Inkscape in general.
Then we are in an agreement. I just singled out bugfixing as the most important aspect of "supporting Inkscape in general", though of course it's not the only one.
For example, my areas of contribution -should- be DOM, binding, and tracing. Yet I never seem to have any time to work on these, with the daily calls for Win32 stuff. I very much want to cut my time on w32 by at least 75%.
So did anything come out of the recent calls to distribute windows lib building?
On Sun, Mar 2, 2008 at 11:18 PM, Bob Jamison <rwjj@...127...> wrote:
But what this upsurge is all about, is people working on their own projects. What I'm talking about is supporting Inkscape in general.
Actually, now that I think of it, this at the same time is a reason for some optimism about the future. I just know it from my own experience. When you are relatively new to the project but already deep into it and feel yourself empowered by the new toy, you're spewing out grand projects and innovative ideas, paying less attention to housekeeping. But as time goes on, you settle down - and then refactoring, ironing for consistency, and general upkeep start to take more of your time. You still feel "married" to the project and want it to improve, but you may not have as much grand ideas as you used to :)
Besides, we do have a lot of relatively recent volunteers that are a great help with housekeeping. Just look at the bug tracker for example. There's an incredible level of activity there, and some of the most active bug triagers don't even have developer access yet.
On Sun, Mar 02, 2008 at 07:59:19PM -0600, Bob Jamison wrote:
Wow, I wish that we could harness all of this email-writing energy on the topic, and focus it on the actual work.
Totally agreed. Our motto early in the project was "Patch first, discuss later."
You were absolutely right on your blog. The ratio of developers to users on win32 is much lower than on Linux. If we had just one or two more developers, it would -double- our abilities.
I am hopeful a silver lining of this will be an increased awareness of our need for help in maintaining the Windows port. We get so many Windows bug reports; I really appreciate the work you've put in to investigate and solve so many of them, but I can definitely imagine how burned out you must feel on this.
Since the win32 stuff has never been a release in its own right, but just an alternate way of building Inkscape, I don't see any problem with releasing Inkscape immediately, and getting a win32 build out whenever it's ready. People can always get a copy off the server. I mean, we build it several times a day. Just pick a build and package it. And if the users don't like it, then we can refund their money. I am a big fan of Fedora, and would -love- to see 0.46 in Fedora 9.
Thanks, yes it sounds like we'll take this approach. The main concerns Bulia and others have voiced around doing this is in relation to the PR, and having a divided message. I'm not as concerned about that but can see the point, and the proposal to delay PR for a few weeks to give the Windows port additional time to catch up sounds like an acceptable compromise, and would maximize Inkscape's uptake on both Win and Lin.
I'll see what patches we need to get printing working well again soon. Maybe we can do the switch to using the new Cairo1.5+ win32_printing_surface in a patch ourselves, rather than wait for Gtk to do it. If so, then maybe a Cairo guy could help us. There ya go. A good opportunity to delegate. :-)
I understand 1.6 is also coming real soon now.
I have been slightly annoyed recently with the vaguely implied idea that we don't -want- to fix this bug, and that any problems we have with Inkscape must be a symptom of our laziness and disrespect for the user. I can assure people that considerable effort has already been spent on this. I have been hoping that the increased clamoring would be accompanied with increased volunteering. I haven't seen it yet, but Hope Springs Eternal.
I was reflecting back on what got all of us involved in Inkscape to start with. By and large I think we each got involved not because Inkscape was perfect, but the contrary - we saw Inkscape's potential, but something was broken or missing and we felt we could fix it.
So, to look for a silver lining in our current Windows troubles, putting a release out without a Windows build could actually work out well in the long term: It could be the call-to-arms that brings in more people to assist in maintaining it.
Bryce
bulia byak wrote:
I disagree. This will defuse the bang.
I think that unfortunately there's never going to be much of a bang on Windows, because it is always just downloading the available file, not coming installed with a new distro/system.
There is a chance for a bang on Linux, maybe twice. The thing is that to make any bang, we need to act quickly for Linux. From a marketing perspective, I'm thinking distro reviews, which tend to itemise at least some of the new apps, will be helpful. There is no such equivalent for Windows or Mac.
As such, I don't mind the bang for Linux followed by less of a bang for Windows, but what I do mind is a strong negative, like severely crippled release versions on any platform.
Is there any way to get the release into the freeze versions of Fedora and Ubuntu on time, but still do the proper Inkscape launch only with the proper launch of the earliest of those distro's? By then all the bugs will probably be sorted on all platforms...
On Sat, 2008-03-01 at 17:36 -0800, Bryce Harrington wrote:
On Sat, Mar 01, 2008 at 10:34:30AM -0500, Gail Carmichael wrote:
So, my question is, what are we looking at for a release date at this moment?
Fedora 9 goes into feature freeze and beta freeze March 4th, and beta release March 13th with Final Development freeze April 8.
Ubuntu Hardy entered feature freeze Feb 14, and goes to Beta Freeze March 13th with beta release March 20th.
For inclusion for both Fedora and Ubuntu, we'd need to release by the 4th. If we skipped Fedora, we'd have until the 13th. If we miss both of those dates, then the next opportunity to get Inkscape out officially to Linux users would be 6 months from now.
Bryce
I would add that windows DOES NOT have release dates / dates for inclusion. Get it into the distros and fix windows as quick as possible and do a . release with docs pointing to how to get the windows release done.
This community rules and doesn't suffer from the problems that other ones do where this approach would mean that windows no longer gets developed...just a situation to deal with rather than precedent.
The other option is for mythical hacking to get windows ready in time... ;)
Jon
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Bryce Harrington wrote:
Fedora 9 goes into feature freeze and beta freeze March 4th, and beta release March 13th with Final Development freeze April 8.
For inclusion for both Fedora and Ubuntu, we'd need to release by the 4th. If we skipped Fedora, we'd have until the 13th. If we miss both of those dates, then the next opportunity to get Inkscape out officially to Linux users would be 6 months from now.
Note that Fedora development (rawhide) currently tracks the Inkscape development (right now pre2) in the good faith that Inkscape will follow the expectations and do a release in due time. So Fedora 9 Beta *will* have Inkscape 0.4 pre2 unless something really bad happens and force a revert.
Also the pre releases are built for the stable branch (F8) and available in the build system (koji) but will be pushed as updates only *after* the formal release.
Being a Windows user, I agree with Bulia on just about everything.
Isn't Linux supposed to have this great package management / update mechanism? (I have used Linux for half a year, Fedora 6/7 and it tells me everytime it boots that there are new packages to install) On Windows, there is no such thing. So when I for example read: "new release!" I go to the website and download; however if we release without a win32 binary, I'd go to the website only to find there is no new release for me. So I forget about it, and then I hear after months: "new release 0.47!" And then I'll see 0.46, but no 0.47 win32 binary yet...
I.o.w. I do not quite grasp what the release dates of linux distros are for, other than that it is all nicely packaged on a CD or DVD; but I am sure most users would check internet for new updates anyway?
Thanks for waiting on a win32 releasable 0.46. (know that there are still *crash* bugs that feature on all platforms and are being fixed aswell)
Regards, Your faithful Windows-only developer,
Johan
J.B.C.Engelen@...1578... wrote:
Being a Windows user, I agree with Bulia on just about everything.
Isn't Linux supposed to have this great package management / update mechanism? (I have used Linux for half a year, Fedora 6/7 and it tells me everytime it boots that there are new packages to install) On Windows, there is no such thing. So when I for example read: "new release!" I go to the website and download; however if we release without a win32 binary, I'd go to the website only to find there is no new release for me. So I forget about it, and then I hear after months: "new release 0.47!" And then I'll see 0.46, but no 0.47 win32 binary yet...
I.o.w. I do not quite grasp what the release dates of linux distros are for, other than that it is all nicely packaged on a CD or DVD; but I am sure most users would check internet for new updates anyway?
it has to do with what goes into the official repositories besides the DVD versions. Thus, users doing an "sudo apt-get install inkscape" or "yum install inkscape" will get the version in the official repository. To avoid dependency problems, only security patched updates get added to the repositories (usually in a separate security respository that is checked in your "update" program).
Nothing stops Inkscapers from having specific packages on it's own website, it's just more work for someone to be a maintainer for each version posted and all the dependent libraries that are not included in a user's installation won't be automatically installed (part of what makes a "great" package management system). Thus, it's much harder for first-time users to install.
-----Original Message----- From: MilesTogoe [mailto:miles.togoe@...400...] Sent: maandag 3 maart 2008 11:07 To: Engelen, J.B.C. (Johan) Cc: inkscape-devel@lists.sourceforge.net Subject: Re: [Inkscape-devel] 0.46 release debate
J.B.C.Engelen@...1578... wrote:
I.o.w. I do not quite grasp what the release dates of linux
distros are for, other than that it is all nicely packaged on a CD or DVD; but I am sure most users would check internet for new updates anyway?
it has to do with what goes into the official repositories besides the DVD versions. Thus, users doing an "sudo apt-get install inkscape" or "yum install inkscape" will get the version in the official repository. To avoid dependency problems, only security patched updates get added to the repositories (usually in a separate security respository that is checked in your "update" program).
Nothing stops Inkscapers from having specific packages on it's own website, it's just more work for someone to be a maintainer for each version posted and all the dependent libraries that are not included in a user's installation won't be automatically installed (part of what makes a "great" package management system). Thus, it's much harder for first-time users to install.
Sorry to get off-topic here but... so the whole official repository is updated *only* a couple of times (say 3) per year with program updates that are not security patches? I find that hard to believe with the updates I was getting. Perhaps I installed a more progressive repository... (because it did install all dependencies automatically etc.) ? How do Linux users get 0.46.1 ? Manually?
Confused, Johan
J.B.C.Engelen@...1578... wrote:
From: MilesTogoe [mailto:miles.togoe@...400...] it has to do with what goes into the official repositories besides the DVD versions. Thus, users doing an "sudo apt-get install inkscape" or "yum install inkscape" will get the version in the official repository. To avoid dependency problems, only security patched updates get added to the repositories (usually in a separate security respository that is checked in your "update" program).
Nothing stops Inkscapers from having specific packages on it's own website, it's just more work for someone to be a maintainer for each version posted and all the dependent libraries that are not included in a user's installation won't be automatically installed (part of what makes a "great" package management system). Thus, it's much harder for first-time users to install.
Sorry to get off-topic here but... so the whole official repository is updated *only* a couple of times (say 3) per year with program updates that are not security patches? I find that hard to believe with the updates I was getting. Perhaps I installed a more progressive repository... (because it did install all dependencies automatically etc.) ?
That will vary from distro to distro and from package to package.
Some distros are very conservative compared with others, some package are more sensitive than others (they may touch security areas or may be dependencies for other applications) and some features are so important that people don't want to miss them.
How do Linux users get 0.46.1 ? Manually?
If the distro ship with 0.46, then one can expect 0.46.1 will get pushed as an update and installed automatically. If the distro ship with 0.45.x then the user will have to install it manually (probably from autopackage and break his update system) or *if* the distro policy allows it and the maintainer want to do the work, as an automatic update.
As we talked about Fedora as a specific example, both the users of the current release (F8, which shipped with inkscape 0.45) and the next release (F9, which most likely will ship with 0.46) will get 0.46.1 as an automatic update. But this will happen for F8 only because the specific Fedora policy allow the maintainer to backport it and the maintainer want to do the job.
Note that it will not be ported for the previous release, F7, which approach its end of life soon (probably June), so will stay at 0.45.1, but one can either install 0.46.x manually from autopackage or from rpm if he can solve the dependencies *or* recompile the srpms.
On Mon, Mar 03, 2008 at 11:16:46AM +0100, J.B.C.Engelen@...1578... wrote:
J.B.C.Engelen@...1578... wrote:
I.o.w. I do not quite grasp what the release dates of linux
distros are for, other than that it is all nicely packaged on a CD or DVD; but I am sure most users would check internet for new updates anyway?
it has to do with what goes into the official repositories besides the DVD versions. Thus, users doing an "sudo apt-get install inkscape" or "yum install inkscape" will get the version in the official repository. To avoid dependency problems, only security patched updates get added to the repositories (usually in a separate security respository that is checked in your "update" program).
Nothing stops Inkscapers from having specific packages on it's own website, it's just more work for someone to be a maintainer for each version posted and all the dependent libraries that are not included in a user's installation won't be automatically installed (part of what makes a "great" package management system). Thus, it's much harder for first-time users to install.
Sorry to get off-topic here but... so the whole official repository is updated *only* a couple of times (say 3) per year with program updates that are not security patches?
Correct, although it's 2 times per year, not 3.
Speaking only for Ubuntu (but I suspect Fedora is similar), in theory the updates can include not only security patches but also high priority, targeted patches for specific crash or data-loss bugs.
I found out all of this the hard way with 0.45.1. They would not accept 0.45.1 for Feisty. I had to submit each individual patch included in 0.45.1, one-by-one, each with a justification, test case for reproduction, and analysis of regression potential. In the end I think I only managed to get a small portion of the patches from 0.45.1 accepted.
Ubuntu also has a "backports" repository, which users in-the-know can enable to get new releases of things. I've no idea if 0.45.1 made it in for that.
I find that hard to believe with the updates I was getting. Perhaps I installed a more progressive repository... (because it did install all dependencies automatically etc.) ? How do Linux users get 0.46.1 ? Manually?
I assume that the vast majority of Linux users just use whatever is included in their distro. The distros make it so easy to use the packaged software that unless you're an extremely hard core user, you're not likely to be in the habit of downloading one-off releases of software.
For Ubuntu, Hardy will be having a 8.04.1 release, and perhaps our 0.46.1 would be included then. Otherwise, Ubuntu users would not see it until 8.10 this fall. I don't know how Fedora works, but imagine it would be similar.
0.45.1 (released slightly after Feisty was released) was essentially unavailable to Ubuntu users until Gutsy was released 6 months later. This was particularly upsetting to me, because ironically I'd invested a considerable amount of time coordinating 0.45.1 in order to benefit Ubuntu, yet it turned out that every operating system *except* Ubuntu ended up including it.
If we were very strict and regular in how we handle patch updates for 0.46.1 (such as requiring test cases or steps to reproduce, regression analysis, and so forth), we may be able to get 0.46.1 accepted as a Stable Release Update in Ubuntu, but the chances of that are essentially nil. However, if there are individual crash or data-loss bugs we feel MUST be fixed in Linux, yes we can get those included given a sufficient amount of paperwork, justification, testing, etc. etc. to assure that there will not be any chance of regressions of any sort.
In other words, in practice the updates are essentially just going to be security updates - anything else is not going to be worth the time effort invested.
Bryce
On Mon, Mar 3, 2008 at 2:09 PM, Bryce Harrington wrote:
0.45.1 (released slightly after Feisty was released) was essentially unavailable to Ubuntu users until Gutsy was released 6 months later. This was particularly upsetting to me, because ironically I'd invested a considerable amount of time coordinating 0.45.1 in order to benefit Ubuntu, yet it turned out that every operating system *except* Ubuntu ended up including it.
Dunno for most everyone else, but I'm starting to feel a little lost in the thread.
Could we possibly have a risks estimation for all cases?
Or, at least,
1) define date of release for all platforms; 2) define date of release to be in time for both FC9 and Hardy; 3) define date of release to be in time for just Hardy.
Do we have any *other* showstoppers left ?
P.S. Please remember that not all of the translations are in the best shape right now.
Alexandre
On Mon, Mar 03, 2008 at 05:58:04PM +0300, Alexandre Prokoudine wrote:
On Mon, Mar 3, 2008 at 2:09 PM, Bryce Harrington wrote:
0.45.1 (released slightly after Feisty was released) was essentially unavailable to Ubuntu users until Gutsy was released 6 months later. This was particularly upsetting to me, because ironically I'd invested a considerable amount of time coordinating 0.45.1 in order to benefit Ubuntu, yet it turned out that every operating system *except* Ubuntu ended up including it.
Dunno for most everyone else, but I'm starting to feel a little lost in the thread.
Agreed, and agreed with Ishmal that it sure would be nice if half the energy that went into this thread was directed at producing patches to fix the issue. ;-)
Could we possibly have a risks estimation for all cases?
Or, at least,
- define date of release for all platforms;
- define date of release to be in time for both FC9 and Hardy;
- define date of release to be in time for just Hardy.
March 13th would be the last possible date for Hardy. I can't speak definitively for FC9, but their beta-freeze is the 4th; they ship beta earlier than Ubuntu but ship their final release after, so the strictness of their beta may be a touch less. So possibly we could go as late as the 13th.
Do we have any *other* showstoppers left ?
There are 8 milestoned bugs remaining, only one of which is critical. Three bugs including the critical one have people working on patches, and need testing feedback. Three are Win32 bugs relating to extensions. The other two have to do with issues when importing SVG files, which I'm not sure really qualify as "showstoppers" and that I think can be pushed to 0.46.1 or 0.47.
P.S. Please remember that not all of the translations are in the best shape right now.
What are the issues?
Bryce
On Sat, Mar 1, 2008 at 7:05 AM, rygle <pittos@...1208...> wrote:
Remember too, that we are mainly talking about one bug, the https://bugs.launchpad.net/inkscape/+bug/179988 ability to print , which IMHO must be working for a Windows release.
I agree. I think the discussion is mostly prompted by Bryce's comment that if we don't right now have someone to work on this bug, we must not let it stop the release. With this I disagree. This might be true of other bugs but not this one. With such a serious regression, we must indeed wait as long as necessary for someone to come up and fix it.
rygle wrote:
Just my two cents worth on the release being held back by Windows;
I think that, regardless of your position on Windows as an operating system, or Windows users, that putting out a release that is crippled would be downright humiliating to the *whole* Inkscape project.
I'm a Windows user. I use Windows daily. I'm very thankful when the OSS software that I use is supported on Windows. But there are plenty of OSS applications that aren't supported on Windows. It just can't happen in the FOSS ecosystem without people. People who are willing and interested in volunteering time and effort to support their platform. I do not thing a Windows-less release is humiliating to the people in the Inkscape project who have worked hard this release cycle to add features and fix bugs and document and translate and triage and advertise. These people have done a great job and deserve commendation. A Windows-less release is only humiliating to the Windows using portion of the community.
Let's ask, what is pushing us so hard to get the next release out? In a lot of these projects there is an impetus based on saving face, which is about bringing new features to fruition. IMHO, the saving face needs to be done by being a little patient, or at the very least not releasing 0.46 for Windows until it is pretty right.
Please explain how you have arrived at the supposition that the motivation for this Inkscape release is about "saving face"? It is a goal of our project to have timely releases. This is a good goal for every OSS project. It helps projects to move forward and gauge progress. And it helps us slow down and focus on identifying and resolving a large number of issues with functionality and stability.
Again, what is the aim? Is it purely to further the name of Inkscape, or is there at least some desire to further SVG as a viable option is a largely proprietary world? I have heard that said a number of times, and used as a reason to include or reject features. This project is meant to be about a good SVG editor, and the rest is a nice by-product. Again, a good reason to hold off just a little.
Please explain how you have arrived at the supposition that this release could be purely to further the name of Inkscape.
Let us also ask, if Linux users are so technically superior, aren't they already using the release candidate, which offers pretty much what the release will? Surely even the nightlies are very stable by now if all the bugs are ironed out. If so, in at least one sense who cares if there's a release if the RC is working and those who want the release are already using it! Linux users are well and truly used to beta and release candidates, and in fact the pure only use the bleeding edge stuff and look down on those who don't. Or are we next going to start railing against Ubuntu users for their lack of technical expertise, and because they need a packaged version? Afterall, isn't Ubuntu trying to attract that Windows-like person?
I find this paragraph insultingly stereotypical. There are technical and non-technical people in every demographic. Linux users very often prefer to use their distro's package management system. When I use linux (Ubuntu), this is my preference and I don't believe myself to be particularly lacking in technical expertise.
Lets please hold off just a little, at the very least on the Windows release. For many Linux users it makes no difference whether or not there's an official release, because the pre2 package is stable. For the Windows users, and the reputation of Inkscape, SVG, and even FOSS, it makes a huge difference.
I believe that the arguments you have made, because of their tone, do more to distract from your argument rather than support it. The lack of a Windows package for this release would sadden me greatly. But I cannot demand that developers who don't have access to windows for philosophical reasons, monetary reasons, or even simple preference support the development for that platform. If no one will support Inkscape on Windows, it cannot exist.
Aaron Spike
rygle wrote:
Just my two cents worth on the release being held back by Windows;
I think that, regardless of your position on Windows as an operating system, or Windows users, that putting out a release that is crippled would be downright humiliating to the *whole* Inkscape project.
Someone not used to the way Linux distributions are done and how a lot of open source projects are handled could get that kind of misunderstanding. I've re-read the whole thread,and unless Google filtered some emails, I can assure you your perception might feel a bit spacey to most people following the discussion.
The open source world has a habit of frequent releases. They sometimes cause regressions, but it's often like that in the OS world. Some features that work in Ubuntu 6.10 stop working in 7.04, and sometimes they only get repaired in the next version - a 6 month wait.
It's not disrespect for the users (although when you're not used to it it can feel like that). It's just that fixing this function would prevent others to use another needed feature, and time being of the essence (esp. with the 6-month-release distros), you can't have your cake and eat it.
On the other hand, Windows program installation is a tad different than what is done on a Linux distribution. Since users grab the program and install it themselves, it makes it possible to release a fix really fast - that is, Windows users would only be hurt for a few days (the time to squash the bug), something most people would consider a slight annoyance compared to having no other choice than to wait 6 month for a fix (which is the case in, for example, Ubuntu).
Up to now the people responsible to get Inkscape out have always cared not to fall into that trap. Each time an updated code would provide bad results in certain cases, the outdated code has been provided to allow people the same behaviour they were expecting to get (like with the two different .pdf export modes).
That doesn't mean everybody thinks the same (and it's ok since it's the final decision that matters), and they're free to say what they think would be the best. Personally, even though I know it won't happen, I wouldn't find disrespectful to release 0.46 on all platform and advertise it, even if print sill doesn't work on Windows. Just advertise it so people know what to expect, and 0.46.1 will fix the bug like other .1 releases have done for other bugs before. That, or just start the campaign for 0.46.1 (see my comments below - who remembers Open Office 2.0 anyway when what we all started to use was at least 2.0.4?).
Let's ask, what is pushing us so hard to get the next release out? In a lot of these projects there is an impetus based on saving face, which is about bringing new features to fruition. IMHO, the saving face needs to be done by being a little patient, or at the very least not releasing 0.46 for Windows until it is pretty right.
Again, what is the aim? Is it purely to further the name of Inkscape, or is there at least some desire to further SVG as a viable option is a largely proprietary world? I have heard that said a number of times, and used as a reason to include or reject features. This project is meant to be about a good SVG editor, and the rest is a nice by-product. Again, a good reason to hold off just a little.
I don't really understand how one could have gotten the idea that releases would be done to save face. Inkscape releases have been opportunities to show Inkscape's features to people that don't know yet about it - a nice habit from a healthy developed program.
The aim has never sounded to me as a face saving option. It is quite normal to desire sharing with as many people as possible the benefits of a program one (not me, I'm not a programmer) has devoted efforts to.
That mean simultaneous releases on all platforms. Which might not be feasible with 0.46 - but which could be done for 0.46.1.
Let us also ask, if Linux users
[rant, snipped]
Prejudices _hurt_ even when you don't take them personally. I don't know how you'd appreciate being told you're a moron because - f.e. - of your skin colour, but remember most people don't like it.
To be clear : most Linux users are on Linux because It Just Works. You pop in the CD, and 20 min later you're free of most hassles. No antivirus to update, not hd to defragment, no freakin' firewall to set up, no looking on the Net for hours to install all the programs you need, no dozens of CD+license keys to install, no hopeless search for an obscure registry key, no freakin' drivers to download - they're on the CD, and if they're not they should be. All the apps are already in the menus, and they update themselves magically when there's security fixes. Most Linux users just don't care for technical expertise because _they don't need it anymore_ - and that's why most switch in the first place. Being told that we're techies, geeks, or anything like that is just a cheap way to single out people.
And to come back to the idea of beta releases for Linux users: as a "Linux user",I hate it when something break, because I'm not used to it. It shouldn't happen on production machines, and I'll never learn to cope with it (or I would be using Windows).
For many Linux users it makes no difference whether or not there's an official release, because the pre2 package is stable.
Pre2 packages are not supposed to appear in any sane distribution. I wouldn't be happy about it, and none would be, whatever the OS, even if you tell me it's stable.
Getting 0.46 final in both Fedora 9 and Ubuntu 8.04 doesn't mean the advertising campaign need to start in March or April. Most users won't upgrade or install their OS right away, and the users that will do so on release day aren't really concerned by the launch advertising campaign (i.e. they will have an updated Inkscape anyway).
However, not having 0.46 in the repositories in time means there will be no way for the vast majority of Ubuntu Inkscape users (speaking about what I know, Fedora might be more reactive) to test and use 0.46. Only a very small percentage of Linux Inkscape's userbase would use autopackage.
Considering 8.04 is a _LTS_ hurts even more. Ubuntu policies being what they are, there's the possibility 0.46 won't get in the repos if it's too late, and only a handful of people use the backports. That's _3 years_ more of 0.45.1, and that's a 3 years nightmare of backporting the security fixes from 0.46 to 0.45.1. That will most certainly mean they'd be happy to SRU 0.46, but the decision will be done by people that probably aren't in this mailing list.
If 0.46 isn't in 8.04, it will never be in the repos at the same time as the advertising campaign - which mean no true "simultaneous release, since an autopackage sitting idly on inkscape.org isn't as useful. From experience, even for updates that fulfil Stable Release Updates requirements, it takes almost 2 months for the packages to get reviewed, approved and released. Not including the annoying fact that usually the package has first to be done for 8.04+1 (8.10),and 8.10 base won't be there till a few weeks after release (mid May?).
It is a somewhat painful aspect of some Linux distributions, and the problem showed itself in the past. If 0.46 can't be ready on Linux in time, it will make it in 6+ month but it already happened last October (got to use Gmail Archives :) ), and it's already the 6+ month version.
On Sun, Mar 2, 2008 at 9:11 PM, Loïc Martin <loic.martin3@...400...> wrote:
would be the best. Personally, even though I know it won't happen, I wouldn't find disrespectful to release 0.46 on all platform and advertise it, even if print sill doesn't work on Windows. Just advertise it so people know what to expect, and 0.46.1 will fix the bug like other .1 releases have done for other bugs before. That, or just start the campaign for 0.46.1 (see my comments below - who remembers Open Office 2.0 anyway when what we all started to use was at least 2.0.4?).
A few years ago, releasing 0.xx.1 was considered an emergency measure and a bit of "losing the face" for the project.
Then, it came about to be a pretty normal and even expected development with most releases.
And now, you're proposing to basically place all bets on a 0.xx.1 and just sweep 0.xx under the rug as some kind of a buggy preview.
I don't think I like this particular line of evolution.
On Sun, Mar 02, 2008 at 10:33:03PM -0400, bulia byak wrote:
On Sun, Mar 2, 2008 at 9:11 PM, Loïc Martin <loic.martin3@...400...> wrote:
would be the best. Personally, even though I know it won't happen, I wouldn't find disrespectful to release 0.46 on all platform and advertise it, even if print sill doesn't work on Windows. Just advertise it so people know what to expect, and 0.46.1 will fix the bug like other .1 releases have done for other bugs before. That, or just start the campaign for 0.46.1 (see my comments below - who remembers Open Office 2.0 anyway when what we all started to use was at least 2.0.4?).
A few years ago, releasing 0.xx.1 was considered an emergency measure and a bit of "losing the face" for the project.
No. A few years ago we lacked the project resources to support more than just one release at a time. We have grown a great deal since then, and supporting dot releases like all other open source projects is now entirely within our capabilities.
Bryce
-----Original Message----- From: inkscape-devel-bounces@lists.sourceforge.net [mailto:inkscape-devel-bounces@lists.sourceforge.net] On Behalf Of Bryce Harrington Sent: maandag 3 maart 2008 13:15 To: bulia byak Cc: Loïc Martin; inkscape-devel@lists.sourceforge.net Subject: Re: [Inkscape-devel] 0.46 release debate
On Sun, Mar 02, 2008 at 10:33:03PM -0400, bulia byak wrote:
On Sun, Mar 2, 2008 at 9:11 PM, Loïc Martin
<loic.martin3@...400...> wrote:
would be the best. Personally, even though I know it
won't happen,
I wouldn't find disrespectful to release 0.46 on all
platform and
advertise it, even if print sill doesn't work on Windows. Just advertise it so people know what to expect, and 0.46.1
will fix the bug like other
.1 releases have done for other bugs before. That, or
just start
the campaign for 0.46.1 (see my comments below - who
remembers Open
Office 2.0 anyway when what we all started to use was at
least 2.0.4?).
A few years ago, releasing 0.xx.1 was considered an
emergency measure
and a bit of "losing the face" for the project.
No. A few years ago we lacked the project resources to support more than just one release at a time. We have grown a great deal since then, and supporting dot releases like all other open source projects is now entirely within our capabilities.
I think we should really avoid 0.xx.1 releases, because Inkscape has no automatic updating or a way of notifying the user that there is an update. It is hard to advocate usage of Inkscape over Illustrator or CorelDraw when it is easy to crash. Inkscape might be more convenient, but losing your work after again a crash is very very annoying, and scares many users away (if it weren't for me standing behind them begging them to use Inkscape...). Which is why I usually only search the bugtracker for the 'crash' tag. I think a user does not expect a new release to be buggier at old features than the old release.
Right now we still suffer from crashes that are easy to trigger (for example with the extensions, I hope live preview is turned off by default?), which is my greatest concern. I hardly ever print drawings, and actually don't even know what Inkscape's print dialog looks like. So please note that there are concerns other than printing, that also count for the Linux release.
Regards, Johan
bulia byak a écrit :
On Sun, Mar 2, 2008 at 9:11 PM, Loïc Martin <loic.martin3@...400...> wrote:
would be the best. Personally, even though I know it won't happen, I wouldn't find disrespectful to release 0.46 on all platform and advertise it, even if print sill doesn't work on Windows. Just advertise it so people know what to expect, and 0.46.1 will fix the bug like other .1 releases have done for other bugs before. That, or just start the campaign for 0.46.1 (see my comments below - who remembers Open Office 2.0 anyway when what we all started to use was at least 2.0.4?).
A few years ago, releasing 0.xx.1 was considered an emergency measure and a bit of "losing the face" for the project.
Then, it came about to be a pretty normal and even expected development with most releases.
And now, you're proposing to basically place all bets on a 0.xx.1 and just sweep 0.xx under the rug as some kind of a buggy preview.
I don't think I like this particular line of evolution.
Point taken for the .1 campaign, because after writing this email I don't feel it's necessary anymore ;), so sorry for the bad idea.
I haven't followed Inkscape development since the beginning, and thus I'm not really in any position to influence Inkscape evolution. All I can say is that a as a user, I trust .1 releases more than full releases, whatever the program :) but I've never been disappointed by _any_ of the Inkscape releases. The project has never been "losing face" and has shown a quality one doesn't always find in commercial applications. There are issues with things like .pdf export and maybe some other stuff, but everybody understand it's related to the format (other OS apps have the same difficulties with .pdf), or to the libraries you're using - because you have always be open about the issues, and whenever we want to get a new release these issues are explained.
As a user, I also understand that if the release matching the advertising campaign (the terms aren't the best) isn't a simultaneous release, the effect will be totally wasted. I can understand what it can mean to a hard-working developer, because even as a user I want the project to get the best recognition it can get.
But there's also the problem of being included in distributions repositories. Fedora users won't understand, and I know I'm also crossed with the policy, but Ubuntu's policies are a *pain* when it comes to SRU. Even a backport wouldn't make it in time for the campaign (unless that campaign starts in mid June), because it not only requires unbelievably tedious procedures, it also requires someone on the Ubuntu side to do their job (it can take a few weeks for someone to notice, and even there the upload can be forgotten). Unless you have an Inkscape supporter in the council, it's a 50-50 risk. And that would mean even more time and energy lost than has been spent on this thread - and only an Inkscape developer's time would be enough. Something I'd rather not see undertaken, because I'd rather Brice spent his time on the project than on something like that (I speak from experience - I've been through it for far simpler patches - and look at the recent aumix fiasco for another example - http://glyphobet.net/blog/?p=140).
What it boils to is that : - starting the 0.46 release campaign without a Windows version is a complete waste - most reviewers will try it on Windows, and the userbase on Windows is big; - starting the 0.46 release campaign without a version in Linux main distributions' repositories is also a waste, since it amount to a 0.46 release without a working Linux version (if you consider that 98% of the Linux users can't and won't use anything different than what their distro provides).
So there's a problem from whatever point of view we look at it. Except on OSX - we should all get a Mac this time :)
Now, if 0.46 appears in the repos before the official campaign starts, where is the harm, as long as it's a real 0.46 (finished on the Linux side with all developers happy and proud of it)?
Nobody's going to use it before the distro is released (leaves till end of April, a good month and a half). And even if the campaign has to be postponed, only a few reviews of the distros will mention a package that isn't in the default install (shame on this distro, though!). Graphic program reviewers won't review it before the campaign starts (or only as a WIP preview, which can be done already).
When 0.46 for Windows is ready for public adoption, the campaign can start as usual - except this time 98% of the Linux users won't be left in the cold for a few month.
Another benefit of having 0.46 already in Hardy's repos (can't speak about Fedora, they seem far more reactive about it) is that whatever patches are released afterwards for 0.46, they'll be included more easily (=less paperwork).
The other solution, but Brice is better placed to speak about it, is to sponsor a beta release for Hardy, as long as he thinks it's possible to convince Ubuntu developers. However, what's the point in calling it 0.45.999 if before mid-March the source can be good enough for 0.46?
I just want to conclude this long email by all my thanks to all Inkscape developers. For years, it's been the best open source drawing program, (can't compare it with Adobe's products, never tried them) and it is, among all drawing programs, the one with the best potential, the one where the development community is the nicest and the most enthusiastic.
I think that a lot of this conversation is missing the obvious solution. It's probably not necessary, but I think that I should remind people: If they really want Inkscape/win32 released immediately, then all they need to do is fix printing. If it's only the same old people who work on this, then we really haven't accomplished anything. We've already gleaned a lot of information about the problem, and posted it on Launchpad, so anyone can use that as a starting point.
There is a new lib bundle on the server, with fresh C++ libs (gtkmm, glibmm, sigc++, cairomm). I built this from source Saturday and uploaded, so there should be no mismatches between C and C++ versions. The "nightly" builds since then have been using it.
bob
Bob Jamison schrieb:
... There is a new lib bundle on the server, with fresh C++ libs (gtkmm, glibmm, sigc++, cairomm). I built this from source Saturday and uploaded, so there should be no mismatches between C and C++ versions. The "nightly" builds since then have been using it.
bob
Bob, does those libs contain the gtk patch from Launchpad http://bugzilla.gnome.org/attachment.cgi?id=99752 ? Thx. Adib.
On Mon, Mar 3, 2008 at 6:32 PM, Bob Jamison wrote:
There is a new lib bundle on the server, with fresh C++ libs (gtkmm, glibmm, sigc++, cairomm). I built this from source Saturday and uploaded, so there should be no mismatches between C and C++ versions. The "nightly" builds since then have been using it.
By the way, do we ship the OpenType fonts with PS outlines support fix in Pango/Win32?
Alexandre
On 2008-March-03 , at 10:09 , Loïc Martin wrote:
So there's a problem from whatever point of view we look at it. Except on OSX - we should all get a Mac this time :)
yeah you definitely should ;) . . . Well I must admit we also have our share of problems too, including for this particular release. No blocker that I can think of right now though. But we get much less testing than on windows so, unfortunately, I am afraid we'll likely need a .1 release to fix the bugs uncovered by wider testing of 0.46.
JiHO --- http://jo.irisson.free.fr/
Loïc Martin wrote:
bulia byak a écrit :
On Sun, Mar 2, 2008 at 9:11 PM, Loïc Martin <loic.martin3@...400...> wrote:
would be the best. Personally, even though I know it won't happen, I wouldn't find disrespectful to release 0.46 on all platform and advertise it, even if print sill doesn't work on Windows. Just advertise it so people know what to expect, and 0.46.1 will fix the bug like other .1 releases have done for other bugs before. That, or just start the campaign for 0.46.1 (see my comments below - who remembers Open Office 2.0 anyway when what we all started to use was at least 2.0.4?).
A few years ago, releasing 0.xx.1 was considered an emergency measure and a bit of "losing the face" for the project.
Then, it came about to be a pretty normal and even expected development with most releases.
And now, you're proposing to basically place all bets on a 0.xx.1 and just sweep 0.xx under the rug as some kind of a buggy preview.
I don't think I like this particular line of evolution.
Point taken for the .1 campaign, because after writing this email I don't feel it's necessary anymore ;), so sorry for the bad idea.
I haven't followed Inkscape development since the beginning, and thus I'm not really in any position to influence Inkscape evolution. All I can say is that a as a user, I trust .1 releases more than full releases, whatever the program :) but I've never been disappointed by _any_ of the Inkscape releases. The project has never been "losing face" and has shown a quality one doesn't always find in commercial applications. There are issues with things like .pdf export and maybe some other stuff, but everybody understand it's related to the format (other OS apps have the same difficulties with .pdf), or to the libraries you're using - because you have always be open about the issues, and whenever we want to get a new release these issues are explained.
As a user, I also understand that if the release matching the advertising campaign (the terms aren't the best) isn't a simultaneous release, the effect will be totally wasted. I can understand what it can mean to a hard-working developer, because even as a user I want the project to get the best recognition it can get.
But there's also the problem of being included in distributions repositories. Fedora users won't understand, and I know I'm also crossed with the policy, but Ubuntu's policies are a *pain* when it comes to SRU. Even a backport wouldn't make it in time for the campaign (unless that campaign starts in mid June), because it not only requires unbelievably tedious procedures, it also requires someone on the Ubuntu side to do their job (it can take a few weeks for someone to notice, and even there the upload can be forgotten). Unless you have an Inkscape supporter in the council, it's a 50-50 risk. And that would mean even more time and energy lost than has been spent on this thread - and only an Inkscape developer's time would be enough. Something I'd rather not see undertaken, because I'd rather Brice spent his time on the project than on something like that (I speak from experience - I've been through it for far simpler patches - and look at the recent aumix fiasco for another example - http://glyphobet.net/blog/?p=140).
What it boils to is that :
- starting the 0.46 release campaign without a Windows version is a
complete waste - most reviewers will try it on Windows, and the userbase on Windows is big;
- starting the 0.46 release campaign without a version in Linux main
distributions' repositories is also a waste, since it amount to a 0.46 release without a working Linux version (if you consider that 98% of the Linux users can't and won't use anything different than what their distro provides).
So there's a problem from whatever point of view we look at it. Except on OSX - we should all get a Mac this time :)
Now, if 0.46 appears in the repos before the official campaign starts, where is the harm, as long as it's a real 0.46 (finished on the Linux side with all developers happy and proud of it)?
Nobody's going to use it before the distro is released (leaves till end of April, a good month and a half). And even if the campaign has to be postponed, only a few reviews of the distros will mention a package that isn't in the default install (shame on this distro, though!). Graphic program reviewers won't review it before the campaign starts (or only as a WIP preview, which can be done already).
When 0.46 for Windows is ready for public adoption, the campaign can start as usual - except this time 98% of the Linux users won't be left in the cold for a few month.
Another benefit of having 0.46 already in Hardy's repos (can't speak about Fedora, they seem far more reactive about it) is that whatever patches are released afterwards for 0.46, they'll be included more easily (=less paperwork).
The other solution, but Brice is better placed to speak about it, is to sponsor a beta release for Hardy, as long as he thinks it's possible to convince Ubuntu developers. However, what's the point in calling it 0.45.999 if before mid-March the source can be good enough for 0.46?
I just want to conclude this long email by all my thanks to all Inkscape developers. For years, it's been the best open source drawing program, (can't compare it with Adobe's products, never tried them) and it is, among all drawing programs, the one with the best potential, the one where the development community is the nicest and the most enthusiastic.
can there be a "stealth" release to Ubuntu, Fedora with idea that any fixes make their release deadlines giving a month or so for the official .46 release to make it in. This seems what they do for Gnome (although that might be an understandable holdout exception). This idea would commit an official Inkscape release within a month.
On Mon, Mar 03, 2008 at 09:55:18AM -0700, MilesTogoe wrote:
Loïc Martin wrote:
Now, if 0.46 appears in the repos before the official campaign starts, where is the harm, as long as it's a real 0.46 (finished on the Linux side with all developers happy and proud of it)?
can there be a "stealth" release to Ubuntu, Fedora with idea that any fixes make their release deadlines giving a month or so for the official .46 release to make it in. This seems what they do for Gnome (although that might be an understandable holdout exception). This idea would commit an official Inkscape release within a month.
This seems like a very reasonable solution. This would allow meeting the deadlines for Ubuntu and Fedora, yet permit Windows additional time to resolve it's bugs.
The risk is that someone would post something to slashdot during the embargo and get interest stirred up prematurely and mess up the PR - which has happened before. But probably a manageable risk.
We should be able to get individual critical fixes (crashes or data loss only) in once 0.46.0 is accepted, even after beta but before the release. There is more paperwork involved, but far less than happens after release.
Bryce
With all due respect as a user, I've read the whole thread and I suffer the bug with printing in Windows too (as a Windows user).
Can be considered a 0.46.pre3 windows release with the official 0.46 release for *nix?, maybe including a disclaimer about the bug with a message "we are working on it" would be fine, remember that the "OSS in windows" users use to be people beware about bugs and some lack of os-related features.
With kind regards
PS: thanks for all the work, inkscape is a terrific tool and is more polished than a lot of "professional" graphic tools in the market
On Mon, Mar 03, 2008 at 06:00:15PM -0300, Jaime Martínez-Figueroa wrote:
With all due respect as a user, I've read the whole thread and I suffer the bug with printing in Windows too (as a Windows user).
Can be considered a 0.46.pre3 windows release with the official 0.46 release for *nix?, maybe including a disclaimer about the bug with a message "we are working on it" would be fine, remember that the "OSS in windows" users use to be people beware about bugs and some lack of os-related features.
Ishmal has been posting devel builds of Inkscape for Windows for some time, and I would bet he'll do similarly for what is tagged 0.46.0 for testing purposes, even if it still lacks some bug fixes. We just would not be marketing it as the official "Inkscape 0.46 For Windows" until those bugs are taken care of, so it would be more of a beta version.
Bryce
participants (22)
-
unknown＠example.com -
Aaron Spike -
Adib taraben -
Alexandre Prokoudine -
Alvin Penner -
Bill Baxter -
Bob Jamison -
Bryce Harrington -
bulia byak -
Gail Carmichael -
Jaime Martínez-Figueroa -
jiho -
Joel Holdsworth -
Jon A. Cruz -
Jon Phillips -
Loïc Martin -
MenTaLguY -
Michael Yinger -
MilesTogoe -
Nicu Buculei -
rygle -
theAdib