Sven Neumann wrote:
> - New RGB->Indexed quantizer

Although this should generally be pretty good and better than the old quantizer, I was hoping to do a nice long tweaking'n'tuning session for this in the 1.3 timeframe, which is where things get sexy. Unfortunately it didn't work out like that because of time matters, so it's pretty much in the untuned state I landed it in. But it's okay.

I really hope that I can get back to it at some point.

--Adam

Hi Sven,
is it time to flame again ?
Instead of that, please try to read _and understand_ the arguments other people have against your decision, or how should one understand :

"Actually a few magazines already know that the next stable release is supposed to be 2.0 for some time already."

And please try a google search for "gimp 2.0" as suggested. My quick looking reveals "gimp 2.0 backend", 16 bit, cmyk, does qualify for pre press ...

At 18:16 18.06.03 +0200, Sven Neumann wrote:

Hi,

Marc kept stating that there are no user-visible changes in GIMP-1.3. I really don't know how much time he spent with GIMP-1.3 ...

See above. BTW: do I have qualified to have an option ?

but I doubt
that he is aware of the amount of changes that have been made. To clarify things a bit and to justify a 2.0 version number for this release, I made a compressed version of the NEWS file as found in the 1.3 tree. So here's a list of (mostly user-visible) changes. I'm sure I still missed quite a few things...

I could comment on every single feature, but if you just reread your own list you should be able to see that there is nothing major beside the huge rewrite which IMHO only justifies the .0 number as in 'beware of totally new misbehaviour' :-)

And Gimp Python was in the Gimp tree before 1.2 was released ...

Hans

What's new in GIMP-1.3
======================

- Extended the concept of linked items to all kind of transformations - Rewritten and much improved text tool (not quite finished yet) - Rewritten and improved paths tool (not quite finished yet) - Modular display filters and color selectors - Popup selectors for brushes, gradients, palettes and the like - Improved session managment
- Fullscreen mode for the image window - MNG saving
- PSD saving
- Allow to create channels from an image's color component - Added color pickers to levels tool for easier color correction - Much improved undo system that adapts to the memory requirements of the undo steps
- Hooks for plug-in debugging
- Converted the API reference to DocBook XML - New colorblindness display filter - Added SphereDesigner plug-in
- Portability fixes for 64bit platforms - Handle large swap files (>2GB)
- Optional menubar in the image window - Added more widgets to libgimp (GimpColorScale, GimpPickButton, ...) - Added a color selector dock
- Added new layer modes (Softlight, Grain Extract, Grain Merge) - Added Gimp-Python
- Improved UI of color adjustment tools - Added GimpSelectionEditor, a view on the current selection - Cleaned up and improved most plug-ins - Support tile cache > 4GB on machines with 64bit long integers - Added support for large files (> 2GB) - Temporary switch to the Move tool when Space is pressed - Added mnemonics all over the place - Adopted the Thumbnail Managing Standard - Added shortcuts to crop layer or image to selection boundary - Improved tool options and made them dockable - Cleanup of brush, gradient, pattern and palette PDB functions - Allow to choose interpolation for individual transformations - New layer mask initialization modes - Scanline conversion (Path to selection etc.) changed to use libart - Improved preferences dialog
- Themeable user interface
- Dockable windows
- New tool icons
- New RGB->Indexed quantizer
- Added Color Erase paint mode
- Added SF-DIRNAME script-fu parameter - Ported to GTK+-2.x
- Rewrittten and improved almost all of the GUI - Cleaned up the core a lot. The app directory is now broken up into subdirectories that define subsystems with defined dependencies - Seperated GUI from core functionality in almost all places - The core object system does not depend on GTK+ any longer - Split up libgimp and libgimpui in a bunch of smaller utility libraries to be used by plug-ins and the core

Sven

On Wed, Jun 18, 2003 at 07:20:13PM +0200, Hans Breuer wrote:

Hi Sven,
is it time to flame again ?

Please, although I am easily at flaming, I do not intend to do it, nor was it my intent to put off Sven, who works _so_ much, nor is it useful to start a flamewar with sven, who invests so much effort into gimp, without ever wanting a thanks.

It was also not my intent to undermine the efforts that went into 1.3. I *know* a lot has been done.

My *opinion* is just that naming it 2.0 is not a good service, regardless of how much work was put into it.

I am convinced that a version number like 1.4 or 1.8 or 1.8 or 1.10 or... should be equally fine to tell people just how much work and effort has been put into the release.

The reason I was relatively upset is that the main argument is always "everybody else has 2.0", and I simply think that is dishonest and not useful to people.

It's not dishonest because sven et. al. were lazy, to the contrary. My deepest respects for the whole rewrite, cleaning up etc. of the source tree. It sure was a lot of work that wasn't honored accordingly.

I can fully understand if sven gets upset now when he interprets my mails as "you didn't do enough to earn a 2.0".

If making it version 2.0 is the only way to honor the hard work, so be it, I don't oppose this.

I just hope that there are better ways to show this, more fulfilling ways.

Please, don't flame (as I almost started to do in my earlier mails), it's of no use and will only give sven the feeling that people don't acknowledge what he does. A disaster that would be.

On 18 Jun 2003, at 23:35, Sven Neumann wrote:

Hans Breuer writes:

"Actually a few magazines already know that the next stable release is supposed to be 2.0 for some time already."

Perhaps I told one or even two people involved in the computer magazine business about it when I tried to get some support for the conference this summer. I'm sorry but I need to sell this conference at the moment and everyone seems flat broke. We really could need some good marketing and instead you guys take this as an opportunity for flames? Please come back with arguments as soon as you have settled down.

Sven, I think you're right to assume that major technical changes, even if they are invisible to the user, warrant major version numbers. I will go even so far as to say that technical changes are the only thing that should drive version number changes.

However, I also agree with those who say that GIMP 2.0 is in the minds of the people as the one that will bring CMYK, GEGL and other such goodies. With smaller apps it wouldn't be so important. Who cares whether Audacity (a fine tool, BTW!) jumps to 1.2 or 2.0? But GIMP is a well known program, and a lot of users know the meaning of its version numbers.

This is all IMHO.

On Wed, Jun 18, 2003 at 11:35:51PM +0200, Sven Neumann wrote:

Perhaps I told one or even two people involved in the computer magazine business about it when I tried to get some support for the conference this summer.

What did you tell them, that gimp-2.0 will be released or that gimp-2.0 will have all those new features people expect for 2.0?

I'm sorry but I need to sell this conference at the moment and everyone seems flat broke. We really could need some good marketing and instead

Who is "we"? A company? You are selling a conference? So the fact that you mentioned the number 2.0 to (maybe) two magazine people means that this version number must be used? Also, who is "we"? *I* certainly don't need any marketing...

I am sorry, but I feel your arguments in favour of 2.0 get more and more confusing, and rather long-winded. Especially if one considers that magazines and websites already published that gimp-2.0 will have cmyk support etc.. and this doesn't bind us in any way, either.

you guys take this as an opportunity for flames?

Please calm down, I more than once told you that I am not flaming. You are working yourself up into something here, really. No flame was intended, just a discussion about the version number.

For some reason you are getting mad at this discussion, not the arguments. Why is this so? Why are you asking for speaking up if you only go mad at people who do? You'd better not have posted anything if you don't want to hear any responses.

[google search]

impression that most the hits for "gimp 2.0" are caused by "gimp" used on the same page as "gtk+-2.0". What exactly do you want to prove by 116,000 hits on google?

Stop putting words in my or other peoples mouth, please. I don't want to prove anything by 116,000 hits. I want to prove that the expection for gimp-2.0 having some major new features like cmyk etc. is there, and searches like "gimp 2", "gimp 2.0", "gimp cmyk" certainly are good indications that a number of people and websites know about the not-at-all secret plans of adding colourspace support and others for gimp-2.0.

Also, if you really want comparison by numbers, than the number of people writing that gimp-2.0 will have cmyk is certainly larger than the number of magazine people you talked to.

And this is no wonder, as this has been mentioned publicly a lot of times.

See above. BTW: do I have qualified to have an option ?

BTW: Yes, indeed you do. What exactly makes you think you don't?

Your reaction, I guess. Asking for responses and then critizising people for responsing at all.

Please don't take it personally. That's the last thing I or others want. I'd be happy with a disucssion about version numbers, and I laid downmy arguments, namely that there are no major features for a major version number, and the added opinion that we don't need new major numbers just because everything else has (becaus thta's just confusing people).

Other people have added that there are great expectations for gimp-2.0, and I think that 1.6 or 1.8 or so would be a nice number telling people "hey there, this was a hell LOT OF WORK!", without destroying all the plans mentioned over the years.

I even think that not having added major new features, but cleaning up the codebase and adding lots-needed bits here and there, is a good thing, as it enables people to start implementing difficult new features such as using gegl without having to add dirty hacks everywhere.

Oh shit, I got one wrong. But wait, I'll put "Image templates" in as a new feature that I forget to listen.

Yes, and "swapfiles > 2GB" as probably a bugfix, not a feature at all, just like 64 bit cleanlyness is a bugfix, and not a feature.

Let's not quarrel around features. If you insist that there are major features qualifying 2.0, I do not really oppose.

However, I *do* oppose "marketing", "but gtk+ has it" and similar arguments. I simply think it's unneccssary.

From what I expected and understood from the Future of Gimp RFC:

http://www.mail-archive.com/gimp-developer@scam.xcf.berkeley.edu/msg03656.html

Gimp 2.0 will indeed have 16-bit per colour value, CMYK, integration with GEGL, etc. If there's still work in this direction, then I suggest making it version 2.0, and keeping the next stable release as 1.4.

Note that a 1.4 version does not devaluate of all the hard work that was put into the Gimp. By all means, a new secondary version in Gimp is always a big deal, and will be very appreciated.

Best regards,

Shlomi Fish

---------------------------------------------------------------------- Shlomi Fish shlomif@vipe.technion.ac.il Home Page: http://t2.technion.ac.il/~shlomif/

An apple a day will keep a doctor away. Two apples a day will keep two doctors away.

Falk Fish

On Wed, Jun 18, 2003 at 11:35:51PM +0200, Sven Neumann wrote:

And please try a google search for "gimp 2.0" as suggested. My quick looking reveals "gimp 2.0 backend", 16 bit, cmyk, does qualify for pre press ...

I did that since Marc suggested and I spent some time with the results. I didn't make up any scientific statistics but I got to the impression that most the hits for "gimp 2.0" are caused by "gimp" used on the same page as "gtk+-2.0". What exactly do you want to prove by 116,000 hits on google?

I did a more specific search:
http://www.google.com/search?q=gimp+2.0+-gtk+-GTK+GEGL+-gimp-developer+-gegl-developer I excluded pages mentioning GTK plus the developer lists. Still 160 hits, some on major sites. The main argument against naming it "2.0" is that 2.0 is already known as "the all new, GEGL-based GIMP with colospace support etc.".

I'm afraid, there would be lots of people asking "where is CMYK support and where is GEGL?" It is very hard to alter such widespread knowledge. It could get a real PITA.

Bye, Tino.

Hans Breuer wrote:

To clarify things a bit and to justify a 2.0 version number for this release, I made a compressed version of the NEWS file as found in the 1.3 tree. So here's a list of (mostly user-visible) changes. I'm sure I still missed quite a few things...

I could comment on every single feature, but if you just reread your own list you should be able to see that there is nothing major beside the huge rewrite which IMHO only justifies the .0 number as in 'beware of totally new misbehaviour' :-)

Here's the point I think we should be focussing on. The GIMP core has been re-written. This means that there have been lots of bugs introduced, and many old ones backed out. If we use a .4 minor version number (or any 1.x version number), we risk giving the impression that this is a stable point-release, a continuation of the 1.2 branch, with stable interfaces, stable internals and the rest.

Since CVS has what amounts to a re-write of 1.2 (as opposed to the modularisation which was envisaged way back at GIMPCon 2000), we would need to be careful that we don't give the impression of stability with an untested program. The new .0 version number says that the program may be unstable, but that it is considered good to go for production work. That is, in my opinion, the case.

That said, much like the earlier discussions over licencing issues, I find the whole issue a pointless waste of time. I do not think that the version number makes a great deal of difference, and I don't believe there will be a public outcry among the GIMP using public just because we use 2.0 without having operation pipes and CMYK.

Personally I wouldn't be averse to calling the current CVS 2.0, and having a 2.2 pretty quickly afterwards (say around Christmas) as was done with GTK+ to say "All known bugs introduced in the 2.0 release are fixed".

The main point, of course, is that version numbering is irrelevant to content, and arguing about it isn't getting us any closer to a stable release, or a usable GeGL. Can we agree that the version number isn't that important, call it something, and get on with writing software?

By the way, what's the current story with PuPUS? Is it abandoned, or will it get released at some stage post-1.3+?

Cheers, Dave.

On Thu, Jun 19, 2003 at 10:47:42AM +0200, Sven Neumann wrote:

Also, if you really want comparison by numbers, than the number of people writing that gimp-2.0 will have cmyk is certainly larger than the number of magazine people you talked to.

And this is no wonder, as this has been mentioned publicly a lot of times.

Sure, I don't wonder, after all that's what we told them 3 years ago.

We've been telling them for a couple of years might be a more precise description (for some definition of we, not including e.g. me :-) The meme has its own life now...

Whenever a new GIMP 1.3 release is announced, people ask when we will finally start to port it to 2.0. There aren't really so many people out there that know about the plans for 2.0 we made three years ago. Don't let google fool you; from the discussions I followed lately, I came to the impression that people expect a GIMP 2.0 release.

Well, the trailing 2s disease is quite widespread among Gtk+ apps. I wish Gtk+ versions changed more often and to more silly numbers, so people wouln't be so tempted to adjust versions of everything else...

Since anyone who've ever seen a 1.3 screenshot (not speaking about running it) must know it uses Gtk+2, I don't think it would be so hard to fight the Gtk+1 confusion.

Anyway the difference with the two confusions is this:

- When I thought 2.0 would be GEGL based, support CMYK, 16bit channels, ... and when 2.0 is released I can't see any trace of these, I'm disappointed.

- When I thought 1.4 would still be Gtk+1 based, and when it's released I see it's Gtk+2 based, I'm delighted.

Regards,

Yeti

Sven Neumann wrote:

Yes, please. But we probably need to get to a point here. GIMP-something.0 sounds pretty weird for a stable release...

I say it's time for a show of hands. My vote is for 2.0, because there are likely to be lots of new bugs and 1.4 makes it sould like a really stable release.

Who's going to keep count?

Cheers, Dave.

On Thu, Jun 19, 2003 at 10:47:42AM +0200, Sven Neumann wrote:

writes:
Marc, you may not need any marketing but I have been trying to raise funding for gimpcon since february. I do believe that we, the GIMP project, could need some more publicity.

Please don't overrate the stuff I said about magazines. All I said was that we are considering to call it 2.0. That's it.

Ok, here's _my_ deal: *If* you say that not calling it 2.0 would cause problems in fundraising, then you simply win... While my concerns were, for me, important enough to mention them (and argue about them), and while the "gtk+ has 2" etc.. style of arguments were not convincing, this one is.

I mean, this version number has a certain importance in my mind. It's enough to disregard most of the arguments I heard so far, but not important enough to risk funding.

I view that then as a neccessary sacrifice.

ago. Three years are a long time and noone will be surprised if we changed our plans by now.

I still disagree on that, people are eagerly waiting for 2.0 for the very features it should have. Unfortunately.

It would certainly raise some interest (which is good)

It would also again raise some interest later when gegl etc. has been integrated. So if you think maximizing interest is necessary.. go ahead.

Your reaction, I guess. Asking for responses and then critizising people for responsing at all.

Marc, please check who is being quoted and answered. Hans asked this.

Well, you wrote "you guys", and I certainly felt non-reasonably criticised from your other mails. Yeah, I guess I can live with that, I won't mention it again ;)

IMO the changes are major
second argument
lately, I came to the impression that people expect a GIMP 2.0

Well, all that really doesn't sound convincing to me, and since we layed down our aqrguments now, we can either agree, or disagree.

need to quarrel about details of the feature list I posted.

Indeed, it will only raise the impression that I think 1.4 hasn't undergone a lot of work.

In any case, if you *need* the 2.0 release for the conference, go on, that's an argument I will readily submit to.

David Neary wrote:

By the way, what's the current story with PuPUS? Is it abandoned, or will it get released at some stage post-1.3+?

You can look back through the archives for my notes on pupus' state. In summary I had to kill it because of lack of time. An early version was up and running, but today I doubt that even a quite complete implementation would be doing much more interesting stuff than a crossbred gstreamer+gegl (except for some of the interactive-image-processing-specific scheduling niceties, I think).

--Adam

On Thu, Jun 19, 2003 at 12:14:30PM +0200, David Neary wrote:

Yes, please. But we probably need to get to a point here. GIMP-something.0 sounds pretty weird for a stable release...

I say it's time for a show of hands. My vote is for 2.0, because there are likely to be lots of new bugs and 1.4 makes it sould like a really stable release.

Alright, this is a real argument for 2.0. It just sounds less stable than 1.4... Taking the "fundraising" argument into account, I'm indecisive. Therefore, I abstain from "voting" (not that I would claim a vote).

Bye, Tino.

On 19 Jun 2003, at 12:56, Sven Neumann wrote:

writes:

Ok, here's _my_ deal: *If* you say that not calling it 2.0 would cause problems in fundraising, then you simply win... While my concerns were, for me, important enough to mention them (and argue about them), and while the "gtk+ has 2" etc.. style of arguments were not convincing, this one is.

We already have problems in fundraising, I can not tell you if the 2.0 would solve them but I had that plan that involved announcing the 2.0 release number plan. If we decide that we stick to 1.4, I'll have to make up a new one.

Can we know what that plan is? Perhaps we can help. I benefitted a lot from the feedback I got on my GIMP for Windows 1.2.3 press release.

I still disagree on that, people are eagerly waiting for 2.0 for the very features it should have. Unfortunately.

Are they? I don't really know what people are expecting from GEGL integration but it will certainly not be another GIMP once this has happened. When GEGL is used, users will probably not notice that the crappy code that provides the basis for pixel manipulations in the current GIMP has been replaced. We should go for GEGL soon after the next release but it will not be a substantial change from a GIMP users point of view. Only if we then add CMYK as a new colorspace and add proper color management functionality, really new features will be available. These enhancements are not provided by GEGL, GEGL only provides a framework that allows to do such changes in a nice and clean way.

From all the people that addressed me and asked for CMYK support, only one so far was able to explain to me what benefits one can get from working in CMYK. All others would have made things worse since they would have attempted to do color separation w/o any knowledge of the inks and paper used to print the result. To get to a point here, CMYK support is IMO a bit overrated. We surely want to add it but we need to do it proper.

You also mentioned integration with FilmGIMP or CinePaint. Well, it seems there is little interest from the CinePaint people, but if you look at the current state of GAP for GIMP-1.3, it seems that we can already provide quite a few of the features that film people keep asking for.

As to the latter, I don't think so, or there wouldn't be a Film GIMP. Cinepaint exists, because it fulfils a clear need.

Yes, there is a difference between what people need and what they think they need.

An example would be resolution: a completely useless measurement of scale, yet all the people in the print graphics business swear by it. I won't tell you how often art directors have asked me what the resolution should be for the web site designs they are making. I always try to educate them, tell them that only the pixels count, but it would probably be much easier if I told them 74 dpi or some such number.

Similarly, working in CMYK is not a technical necessity: it's a market space demand (although I personally would not mind having blackness as separate channel, but then preferably in a RGBK format).

So you have to ask yourself: who am I selling to? Graphics artists? Geeks? Buyers for large firms? Reporters? The Slashdot crowd? Governments? They all have different needs, and these needs may not be fulfilled by a pretty version number, or by features, or by technical prowess and progress.

If you're trying to sell GIMP progress by organising a meaningful GIMPcon, perhaps asking for money on Slashdot would be more useful than talking to one or two journalists. I don't know. What are your expectations? Does your experience tell you they will come true?

On Thu, 2003-06-19 at 12:14, David Neary wrote:

Sven Neumann wrote:

Yes, please. But we probably need to get to a point here. GIMP-something.0 sounds pretty weird for a stable release...

I say it's time for a show of hands. My vote is for 2.0, because there are likely to be lots of new bugs and 1.4 makes it sould like a really stable release.

I couldn't have said it any better. A new road map (or release plan) will have to be written no matter if we call the next release 1.4 or 2.0 - and changing a three year old release plan is something that most people involved with software will understand.

I say we go for 2.0 - and, as mentioned elsewhere on this list, plan for a quick 2.2 bug-fix release later this year.

Sincerely, ./Brix

On 2003-06-19 at 1504.01 +0200, Henrik Brix Andersen typed this:

On Thu, 2003-06-19 at 12:14, David Neary wrote:

Sven Neumann wrote:

Yes, please. But we probably need to get to a point here. GIMP-something.0 sounds pretty weird for a stable release...

I say it's time for a show of hands. My vote is for 2.0, because there are likely to be lots of new bugs and 1.4 makes it sould like a really stable release.

I couldn't have said it any better. A new road map (or release plan) will have to be written no matter if we call the next release 1.4 or 2.0 - and changing a three year old release plan is something that most people involved with software will understand.

I say we go for 2.0 - and, as mentioned elsewhere on this list, plan for a quick 2.2 bug-fix release later this year.

is this you volunteering to fix the wiki?

http://wiki.gimp.org/gimp/

carol

On Thu, Jun 19, 2003 at 12:56:03PM +0200, Sven Neumann wrote:

I still disagree on that, people are eagerly waiting for 2.0 for the very features it should have. Unfortunately.

Are they?

I do. Others on this list do. It's up to you to make your opinion on that.

I don't really know what people are expecting from GEGL

CMYK, floating-point, programmable layer modes, dynamical effects, layer trees...

happened. When GEGL is used, users will probably not notice that the crappy code that provides the basis for pixel manipulations in the

This from the person who says the gtk+2 rewrite is the major feature people are expecting. Woaw.

From all the people that addressed me and asked for CMYK support, only

one so far was able to explain to me what benefits one can get from working in CMYK. All others would have made things worse since they

Well, it's not just CMYK of course. I am also of the opinion (that I mentioend quite a lot of times), that working in CMYK is not at all the problem, but interoperability is the key problem. postscript paths (For clipping), and cmyk _bit_ format in files (because many programs intrepret rgb as "CMY" or worse).

You also mentioned integration with FilmGIMP or CinePaint.

Never did I use these words! I believe I didn't even quote them. ;) Who is "you", in this case, again?

That said, I don't think I can ensure you that we need 2.0 for the conference but I am still convinced that the amount of added features is worth it.

*sigh*, I am confused. Well, I offer you to decide wether 2.0 is worth it from the fund-raising standpoint, and still state my opinion that 2.0 is a disservice to gimp users, and no service to anybody except maybe a ego push because so much work went into it.

This release will definitely mark a new era of GIMP.

Well, it's exactly as was planned for 1.4 before.. and I really fail to see the new era. My honest aplogies, but that's how I see it. We can rest it here if you want, and agree to disagree. If you agree, I'll be quiet, since I then said all that was to say from my side.

When, if not now, do you want to increase the major version number?

When there is a major change (e.g. gegl, cmyk). Using another toolkit is not a major change at all to me. Using the same internal representation for images, having the same features, simply doesn't warrant the new major number.

I mean, all the concepts in gimp-1.3 are the same as in 1.2, no user visible major changes (yes, lots of small user visible improvements, but none of them qualify as major change). I simply don't think that 100 small improvememnts are one major improvement.

In addition, arguments like "but others have bumped their version number" sound so extrenely fishy and dishonest to me that if such arguments are brought forward as the main and principle arguments to bump the version, I think there is ample reason to question them. "The others do it" is never ever a sound or reasonable argument to me.

I hope the latter paragraph explains why I am opposed so much. It simply sounds fishy to me.

Yet again, I let you decide wether it's important enough for the fundraising issue. *That* is an ugly and difficult to digest argument, but it concinves me.

On Thu, Jun 19, 2003 at 02:03:50PM +0200, Branko Collin wrote:

So you have to ask yourself: who am I selling to? Graphics artists?

(off-topic philosophical rant, not meant as an answer to you!)

Personally, I didn't write gimp-perl (the only major contribution of mine to gimp) to sell it to anybody.

I wrote it to fulfill a need. My need. This need can either be practical (as was in this case), or philosophical (I wrote a zipcracker once simply because I couldn't find a fere one), or being asked by people ("I can do it, and so many people want it"). Another way to describe that as "the need for personal pleasure" or "masturbation" or whatever you want to call it.

Some people might want to "sell". I don't. And everybody who tells me "marc, you have to sell it to the people" will get my sincerest flame, since I write it, you either take it, or leave it. You can comment, or help, or criticise. But if somebody tells me that "you have to xxx" I cna get rather angry, as all this "you want to sell" just presumes what I want or even tries to order me around.

I know not everybody thinks that way. Alan Cox probably thinks similarly, Linus doesn't. There are all sorts of people.

And not all of them feel that "selling" something they wrote is a must, or a need, or even useful.

On Thu, Jun 19, 2003 at 12:14:30PM +0200, David Neary wrote:

I say it's time for a show of hands. My vote is for 2.0, because

My vote is for 1.x, or 2.0, if sven decides it on the grounds that we need it for marketing. The other arguments simply don't overweight the confusion I anticipate. How *do* you count? ;=>

And, actually, I think voting is not useful... we'll have to convince the people with the power (which includes Sven) to do it. Whoever does the release decides. Anarchy. I like it.

there are likely to be lots of new bugs and 1.4 makes it sould like a really stable release.

Just like 1.0 and 1.2, eh? really stable releases, eh? or kernel-2.4, or..

I am sorry, but there are no stable and unstable branches. 1.2 or 1.4 or 2.0 have nothing to do with stability, but all with branching. You expect stability after lots of testing from users, who will not test cvs snapshots.

That is, you create a 1.4.x or 2.0.x branch. This is how it handled about anywhere else, including older gimp versions. changing this wlel-established way if handling releases is going to give much more confusion.

Basically, why don't we just use revision numbers from cvs, or a simple counter... really, the current trend makes version numbers less and less informative, so why keep them at all?

* Sven Neumann [030619 17:18]:

Hi,

writes:

When, if not now, do you want to increase the major version number?

When there is a major change (e.g. gegl, cmyk). Using another toolkit is not a major change at all to me. Using the same internal representation for images, having the same features, simply doesn't warrant the new major number.

OK, so replacing the approx. 8,000 lines of code in the base directory with GEGL would be considered a major feature. The fact that the other 230,000 lines of code that make up the application have been substantially rewritten counts as a minor improvement only? Sorry, I cannot follow you on this argumentation.

The only valid argument for not bumping the major version is the fact that it earlier has been announced that it will happen after some named features and technologies have been integrated. Whether these technologies represent a big change or not is irrelevant, in an continusly evolving application like Gimp quantum leaps of functionality is unlikely.

When GEGL is initially integrated with Gimp no additional features will probably be seen by the user, thus no quantum leap,. after that (I hope), more features og Gimp will take advantage of the approach GEGL takes to image processing, but again no quantum leaps,. is this an

/pippin

My point in this situation is that, given that version number doesn´t matter all that much, why can´t we just be honest with ourselves and call it 1.4? Look, nobody cares about version numbers anymore. Let´s take the Linux kernel as an example: version 2.2 got a lot more of media attention than 2.0.

Sven, if you want to raise funds for the GIMPCon, why don´t you tell your magazine friends that GIMPCon will be the place for the exclusive world premiere of the new stable version of the GIMP? Then you, or anybody else, can plan and write a cool lecture for the presentation. And after it, it can be made available for download worldwide or something like that. You can also write a long article for them about the Convention. They usually pay for that. There are a lot of ways to make money without compromising the integrity of a project.

A new road map (or release plan)
will have to be written no matter if we call the next release 1.4 or 2.0 - and changing a three year old release plan is something that most people involved with software will understand. I say we go for 2.0 - and, as mentioned elsewhere on this list, plan for a quick 2.2 bug-fix release later this year.

I think this is wrong in terms of marketing strategy. Version 2.0 must be reserved for a especially stable and groundbreaking realease. If the long awaited GIMP 2.0 is inmediatelly followed by a bugfix, what would you think of the developers? I can tell you what I´d think: That they were only worried about the "millenarism" of the number and that they probably wanted to make The GIMP look mature at the eyes of non-expert users. That´s what I would say. And reading what some of you think about the issue, I guess I would be right.

In Spain we have a saying that suits this situation: el buen paño en el arca se vende. It means that the perfect marketing strategy is a good and reliable product.

José Manuel García-Patos Madrid

On Thu, Jun 19, 2003 at 05:09:57PM +0200, Sven Neumann wrote:

OK, so replacing the approx. 8,000 lines of code in the base directory with GEGL would be considered a major feature.

If we get all the other stuff we said would be in 2.0, yes.

The fact that the other 230,000 lines of code that make up the application have been substantially rewritten counts as a minor improvement only?

Well, from a user perspective, the improvement from using gtk2 over gtk1 is very nearly nil.... Even for me, the switch from gtk2 to gtk1 in itself is not at all an important new feature or improvement.

On Thu, 19 Jun 2003 10:03:55 +0200 David Neary wrote:

Since CVS has what amounts to a re-write of 1.2 (as opposed to the modularisation which was envisaged way back at GIMPCon 2000), we would need to be careful that we don't give the impression of stability with an untested program. The new .0 version number says that the program may be unstable, but that it is considered good to go for production work. That is, in my opinion, the case.

What ? You mean that Gimp 2.0 won't give the user what you have written and more of that won't be stable ? It's not serious... This will be very confusing for people who explain for years and years that gimp 2.0 will be ... what you have written, and for other GimpPeople. Who will trust GimpPeople now ? You should really considerer that the Gimp Community will loose some credibility.

Furthermore this probably will give lot of arguments to Gimp detractors, why they could not say now that Gimp2 was only born to make definitvly obsolete the_gimp104_fork_that_nobody_here_know_it_exists and no other good reasons ?

Sorry for my english,

@+ Raymond

Hi,

Le jeu 19/06/2003 à 10:55, Sven Neumann a écrit :

You can get 160 hits on google for whatever statement you would like to make. It is a stupid attempt to try to prove anything with a search engine that has billions of pages archived.

I also agree that coming up with a google search is ridiculous.

I'm afraid, there would be lots of people asking "where is CMYK support and where is GEGL?" It is very hard to alter such widespread knowledge. It could get a real PITA.

As I said in another mail already, I believe that it will be a major PITA to explain why it GIMP uses GTK+-2.x and still is not called 2.0. I could surely come up with a google search to proove this but this is getting ridiculous. Anyone who followed the discussions on various sites that announced GIMP-1.3 releases lately, will have noticed that people keep asking for a GIMP port to GTK+-2.x. Going for GIMP 2.0 will IMO be less confusing than sticking to 1.4 just because we stated so 3 years ago.

The problem is not that 2.0 == Gegl was stated 3 years ago, but that everyone (ie: magazines, websites) stated it since these 3 last years.

As a Gimp user I am particullary enthusiastic with the changes introduced in the 1.3 branch, but I think in this scope it would make much more harm to call it 2.0 than 1.4, even if it is based on GTK+2.

But don't be affraid, whatever your decision will be, ppl here will support you. And ppl are confident in you, and the way you have (will) managed the Gimp since all this time.

If you (and the other main gimp developpers) finaly decide to call it 2.0, I think you should really insist, in your press release, on the fact that this realase was so good that you finally decided to call it 2.0 instead of 1.4. And that 2.0 will be 3.0. This would be a really positive and understable way to explain the change.

Thanks,