<Glazblog/>

Kompozer

Kompozer is a fork of Nvu 1.0, fantasai and Jean-Charles Moriaud have mailed me about. Bug fixes here and there, little UI changes. So here's a little clarification:

  • no, Kompozer's lead never let me know about it. I just found a post in nvu forums hosted by geckozone he did but he never had the politeness to let me know about it. Since I am busy and rarely dive into fora - whatever the forum - I was not aware of Kompozer. But the guy is not far away, here in France, and we speak the same language. How hard, and well, sorry to say, how lame.
  • No, I don't think the guy contacted the copyright holder, Linspire Inc., either and that's quite unpolite too even if he has the full right granted by the license to do what he did.
  • no, I don't really appreciate when he calls Nvu bugs "glazbugs". I take that as an insult.
  • yes, his work is a good idea, but no, I don't think this is going to be a very useful mid- to long-term investment since the nvu 1.0 codebase is scheduled to die and be replaced by a xulrunner-based application anyway. It's gecko1.7-based and that's too old.
  • yes, I am working on that new app

Open-source and "free" licenses are not a known excuse for tactlessness.

Comments

1. On Friday 15 September 2006, 02:20 by Klint

Yes, I agree with you globally, but what about those ennoying bugs in NVU ? A minor bug-fix release would be very welcome, even if it's an old platform ! Firefox has just delivered 1.5.0.7 even if 2.0 is expected in a couple of months.
All this being said, it's easy to speak, and hard to do ;) and you already have done a great job with NVU 1.0
But, you know, the ramson of success !!

2. On Friday 15 September 2006, 03:11 by Kazé

Here's my point of view:

#1: I'm working on Nvu extensions since the 0.7 version. I've proposed you a few patches but never got any reply (except a "rejected" in this blog). I've also proposed you my help to integrate Tidy in Nvu.

#2: I *did* try to contact Linspire to propose them some patches, but never got any answer either. Jerry Muelver (nvudev.org forum moderator) has more impact on them than me, but never got any answer either. However, he's been allowed to open a "KompoZer" board on the official forum.

#3: If you've read «glazbug» in the Geckozone forum, you've also read «kazbug». It's just a quick way to separate existing bugs from regressions introduced by KompoZer; the three most active contributors in Geckozone aren't developpers (Do-IT, Ymai, chinon37), the word 'regression' could be misinterpreted.

#4: Last year in Lyon (JDLL 2005), we talked about the difficulty to backport Nvu's code to Mozilla (Flore and Bobo were there). I proposed you to adapt Nvu's code to XUL Runner, you told me that «It's trivial and useless». This is the only reason why I don't work on a XUL Runner version: it would be a useless work since *you* decided to work on it.

#5: Glad to hear it. If you read more about KompoZer, you'll find out that I clearly announce that KompoZer will be obsolete as soon as Composer2 comes true. This is even explained in the KompoZer board on the nvudev.org forum.

I've started this project because many bugs are much easier to correct by patching the source instead of developing extensions. Besides, most users can't even find/install such extensions (remember what you've written about GIMP?). As a moderator of an Nvu support forum, it takes me less time to work an a bug-fix release than answering the same stuff to users that get disapointed by Nvu.

Last but not least: with Jerry Muelver we found that it would be the best way to say that Nvu is still being developped. Many users write that they drop Nvu because of the known bugs that aren't patched.

3. On Friday 15 September 2006, 03:13 by Kazé

By the way: KompoZer should be written with an uppercase Z. ;-)

4. On Friday 15 September 2006, 04:41 by Daniel Glazman

KaZé: no, let's be precise here, in Lyon I said that making a xulrunner-based version of Nvu 1.0 would be non-trivial and useless. I said non-trivial because Nvu still relies on XPFE, and I said useless because if you move to Firefox toolkit, a complete revamping of the code makes sense.
It's not because I rejected some of your contributions that all contributions will always be rejected !!! What a strange reaction.

5. On Friday 15 September 2006, 04:45 by Nuer

Release Early, Release Often ...

6. On Friday 15 September 2006, 04:48 by Daniel Glazman

Nuer: yeah, but don't forget (a) have a private life (b) be paid so you're not starving to death

7. On Friday 15 September 2006, 05:20 by Kazé

Daniel: I remember that very sharply because I stopped my holidays and went to the JDLL mainly to meet you and propose my help. Let's say I don't remember what you said ("just grab XUL Runner and link the libeditor, you're set: this is trivial but useless"), that's no question any more since you're working on the XUL Runner version.

However, this isn't the point: you didn't reject "some" of my contributions, you rejected *all* of them. You never replied *any* of my e-mails. Linspire (Tom Welch) replied to one of my e-mails (the one announcing my first extension), but never replied to my patch proposals either.

About my «strange reaction»: I'm glad I finally deserve your attention, but I won't wait for your blessing to publish my patches! Again, KompoZer isn't a fork since there hasn't been any release of Nvu since the 1.0 version (2006-06-20). For users, it's an unofficial bug-fix release; for Linspire and yourself, take it as a patch proposal.

I don't think I'm doing any harm to the Nvu community with this project. KompoZer follows the exact same tri-licence as Nvu, so that all my patches can be backported to Nvu any time. If you don't have the time, I can do it myself.

8. On Friday 15 September 2006, 05:37 by Daniel Glazman

KaZé: Nvu has a goal. It's not a catch-all-patches project where all contributions are always accepted, becoming a gigantic mess. Like it or not.
I did not accept your site manager patch because we did not want Nvu to become a focal point for object management. I did not accept your URL rewriting thing because I had the feeling a chrome solution was not the good one. And so on.
So you're right about doing a fork. Anyway, the axe will eventually fall on the Nvu 1.0 codebase and probably on Kompozer with it.

9. On Friday 15 September 2006, 05:58 by Kazé

Daniel: I don't get your point of messing Nvu's code, since you're revamping the whole code for the next version anyway. There is already a second trunk for Nvu/Composer2/whatever, so why not patching the old branch if it eases the use and the support?

Anyway, I'm glad I finally get answers about (some of) my patch proposals. Maybe you could give a look at KaZcadeS, too. There must be another good reason not to accept this one.

10. On Friday 15 September 2006, 06:45 by jcmoriaud

Ahem, désolé d'être (un petit peu) à l'origine du fight.
J'avais en effet envoyé ça à Daniel en Août:
"Bonjour, j'ai trouvé ça dans le pourtant excellent site Libellules.ch: www.libellules.ch/dotclea...
C'est quoi ce komposer?
Merci pour NVU!"
Utilisateur de base de NVU que j'apprécie, je n'avais jamais entendu parler de KompoZer. Je me demandais pourquoi apparaissait un dérivé alors qu'il n'y avait aucune information préliminaire. Je comprends maintenant.
Je déplore qu'il y ait une division (fork) plutôt qu'une mise en commun des talents. Un petit peu moins de rigidité et de suceptibilité, un peu plus d'écoute et de collaboration dans le monde du libre seraient bienvenus!
Amen!

11. On Friday 15 September 2006, 07:01 by Kazé

Encore une fois, je préfèrerais de loin bosser sur Nvu (même s'il est destiné à être remplacé) ou sur son successeur que de proposer mes patches tout seul dans mon coin. J'ai proposé mon aide à plusieurs reprises, dont aujourd'hui.

Je n'ai aucune rancoeur vis-à-vis de Daniel ou du projet Nvu, c'est juste que je préfère mettre en application mes patches que d'attendre une éventuelle réponse. Je ne cherche pas non plus à me faire de la pub sur ce projet, mon nom n'apparait même pas sur mon site.

Tous mes patches sont dûment documentés, et la licence permet leur intégration immédiate dans Nvu... en attendant le successeur, que je serai parmi les premiers à télécharger, évidemment. Ca permettrait de montrer aux utilisateurs qu'ils ne sont pas oubliés, et que Nvu est toujours développé.

12. On Friday 15 September 2006, 08:16 by Gérard Talbot

Mr Glazman,

Je fais activement la promotion de Nvu depuis plus d'un an, et ce, de plusieurs façons. Je travaille depuis plus d'un an en étroite collaboration avec Charles Cooke.
Je déplore toutes les chicanes de clochers, les problèmes de communication qui aliènent les collaborateurs, les développeurs et les partisans du code libre. Je suis choqué d'apprendre de Fabian que

"If Daniel and/or Linspire Inc. had accepted my work, there would be an Nvu 1.1 version instead of a KompoZer project."

C'est ce genre de nouvelles qui déplait énormément à tous ceux et celles qui désiraient voir Nvu progresser et s'améliorer, en particulier, voir la sortie de Nvu 1.1 corriger tous les bogues de fonctionnement ou d'implantation que Charles Cooke et moi avions rapportés.

J'ai été un des premiers à avancer l'idée d'intégrer Tidy dans Nvu:

"Personally, I'd be happy if Tidy was built-in and customizable in
Nvu 1.1. I use HTML Tidy with Firefox thanks to Marc Gueury"
July 15th 2005, [Nvu Info List]
en réponse à Dave Sergeant qui mentionnait la nécessité d'utiliser Tidy avec Nvu.

L'intégration possible de HandCoder 0.3.4 dans KompoZer 0.77 et/ou Nvu 1.0 a été certainement une bonne nouvelle, une *très* bonne nouvelle.

Gérard Talbot

13. On Friday 15 September 2006, 08:29 by Daniel Glazman

Gérard Talbot : non vous n'êtes pas le premier ; le premier c'est Tim Berners-Lee. J'ai d'ailleurs une version de développement ici qui intègre Nvu mais sans gérer ses nombreux flags. Nvu2 le permettra.
J'ai ***toujours*** dit et écrit qu'il n'y aura pas - venant de moi en tous cas - de v1.1 de Nvu tant que le code ne sera pas mergé avec cvs.mozilla.org. La divergence profonde entre gecko1.7 et 1.8 a rendu cela quasi impossible. J'en ai tiré la conclusion logique qu'il fallait faire autrement et laisser tomber le code actuel.
Quant aux features, Nvu est le succès qu'il est non pas malgré que je refuse des features, mais bien PARCE QUE je refuse des features. C'est parce que j'ai constamment, avec Linspire, évité les dérives de geeks que Nvu a su se faire une place dans des écoles, des universités, des entreprises, et même, oui même s'ils ralent chez les geeks.
Je n'ai __aucunement__ l'intention de changer cet état de fait. Tout projet a besoin d'une "direction". Je ne prétends pas au titre de "directeur" mais j'assume totalement ma volonté de ne pas voir Nvu devenir le fatras übergeek que Mozilla Seamonkey était devenu et qui a poussé Firefox.
Handcode est un bonne nouvelle pour 0.5% des usagers de Nvu ! Quand vous aurez compris cela, vous aurez enfin saisi que cela DOIT rester une extension et ne doit pas être intégré.

14. On Friday 15 September 2006, 08:49 by jpbardiau

La véritable avancée vis-à-vis du composer de Mozilla ce n'est pas Nvu, c'est KompZer car là il y a eu du travail.

15. On Friday 15 September 2006, 08:53 by Daniel Glazman

désolé, mais alors là, la méga-insulte de jpbardiau ci-dessus me reste ***TRES*** nettement en travers de la gorge. J'ai bossé comme un MALADE full-time entre juin 2003 et juin 2005 pour partir de Mozilla Composer et aboutir à Nvu. Cela a représenté une avancée énorme, qui a permis à Nvu de dépasser largement les deux millions d'usagers.
jpbardiau, soyons clairs : vous êtes une crême d'andouille à un pourcent de matière grise et vous n'êtes plus le bienvenu ici ; tout post de votre part sera instantanément censuré. Allez au diable. A bon lecteur...

16. On Friday 15 September 2006, 10:27 by Kazé

jpbardiau: comme je viens de l'écrire sur le forum Geckozone, le travail que j'ai effectué pour KompoZer n'a rien de comparable avec celui fait par Daniel sur Nvu.

J'essaye au contraire de modifier le minimum de code pour rester "éligible" comme patch de Nvu 1.0. Je n'ai ajouté aucune nouvelle fonctionnalité, à part une option pour nettoyer les URLs locales.

17. On Friday 15 September 2006, 14:03 by alias

Bah,

je trouve ça plutôt sympa, quelqu'un qui se fatigue à corriger les bugs des autres sans essayer d'en tirer gloire. Je ne comprend pas très bien pourquoi cela t'ennerve Glazou : son boulot est pourtant plutôt un hommage à ton travail, non ?

Quand à la provoc de jpbardiau, je n'imagine pas un seul instant qu'il puisse le penser réellement. C'est juste pour t'ennerver et il semble qu'il ait réussi. Fait attention à ta tension quand même ;-)

18. On Friday 15 September 2006, 18:34 by Jason Krew

Could you please fix the ctrl-end bug?

When editing the code in a large html file (eg ~1000000 or greater) ctrl-end never works in going to the end of the code, unlike every other text editor in existence. Like this file for example: kibosig.googlepages.com/f...

Anyway, thanks for the great program!
Now if only someone would code a perfect frontpage to xhtml covertor. Maybe I should learn programming and do it myself so I could clean up my (ugly) site:
www.medi-vet.com/

19. On Saturday 16 September 2006, 17:22 by Chris

I really appreciate the KomPozer fork because I love NVU, but can't be forced to deal with it's many bugs while waiting an unknown period of time for Daniel to come out with composer 2.

I think Daniel should consider releasing at least one bug fix release of NVU before Composer 2 to help people like myself.

20. On Saturday 16 September 2006, 21:43 by JoeS

Well, I thought that was what 'open source' was about. If you don't like the path that the main developement is taking, then make your own path. With all due respect, NVu, Composer, and mailnews html compose seems to be at a standstill. What's left than to 'do it yourself'. Some folks are competent to do that, and I applaude their efforts.
Regarding Gecko composer/Editor I think it is time to shit or get off the pot.
Confirmed bugs that are 3 or 4 years old speak for themselves. As far as mailnews html compose goes, if it can't be supported, (as demonstated by the lack of bug fixes) then just pull it out of the platform, and be done with it. Maybe someone will write an extension to restore basic N4.x functionality. I won't hold my breath.

21. On Sunday 17 September 2006, 04:32 by Martijn

With all respect: please look at what a user would expect from Nvu:
- Nvu 1.0 is a good application, but does not live up to some of its competition. It is a good start and now should move forward. That's what opensource is about.

- Currently we have Composer and Nvu, but it seems that both of these applications cannot seem to go forward with their development because the changes should be merged in some way.

- Where can a user currently learn about Nvu bugs and what's the status of them ? Through Mozilla's Bugzilla ? Through the Nvu forums ?

- Because there are no updates to Nvu, someone finally steps up and does something about this. That effort should be applauded, not critisized.

It would be nice if the author of KompoZer would have contacted you about this.
It would also be nice if you would follow the Nvu-forums to see what the Nvu-users are experiencing on a daily basis.
I expect that the author of KompoZer would have thought that announcing it on the Nvu forums would be enough. One would expect that the Nvu author also reads what's going on on its forums. It is quite stupid that I have read about it on the Nvu forums about a month ago and that only now you have learned about this.

Of course you are busy; we all are. Perhaps some moderator of the forums could give you a brief update when something on the forums should be brought to your attention.

22. On Sunday 17 September 2006, 04:33 by edwood

This is very sad. What kazé did was precisely what makes open source shine. Take an unmaintained branch of a project and do what the license allows you: modify it, make it better, make it meet your needs.


Daniel, you did not know about KompoZer because you lost interest in the community that believed in Nvu. So you do not like "fora" and have to spend time earning your money. Nothing wrong there. But is not good to get angry because the community took charge of the abandoned code, in the absence of response from developer "main land".


Thanks for Nvu.


23. On Sunday 17 September 2006, 11:07 by Daniel Glazman

edwood : you must be kidding. How hard is it to send an email "hi, my name is blabla and I started a fork of Nvu to incorporate bug fixes ; just wanted to let you know" ???