Recent blog entries for uwog

My blog moved here.

21 May 2004 (updated 21 May 2004 at 14:07 UTC) »

Mono

Just my point of view on this whole issue. Disclaimer: I'm a total retard with respect to patent/legal stuff, and I do not work for any company.

It seems to me that the chances of getting in 'trouble' when I would be developing on/using Mono seems to be a lot bigger than when I for example hack on my pet project libwpd. Corel here ensures me that I won't run into trouble as long as I add the following disclaimer: "Note that libwpd is not manufactured, approved, or supported by Corel Corporation or Corel Corporation Limited". I can live with that.

The day I get a letter from any big company located in Redmond related to some currently imaginary OSS project I'm working on, I'll stop hacking on it immediately. I don't have a company backing me up like Nat, Havoc and others have. Keep in mind that there are still 'stand-alone' hackers out there working on GNOME (related things). We can't afford being sued.

I want to concentrate on having fun producing (usable) code; I do not want to run into potential legal problems, and it seems to me that the Mono path is more dangerous than a project like Abiword or libwpd. If only MS would say they'd let us have our fun as long as we tell the world: "Note that Mono is not manufactured, approved, or supported by Microsoft Corporation".

Lucene

Nat: Lucene itself is actually quite nice indeed. I wrote the search functionallity using Lucene for the Content Management system that ASML (the company who makes lithography systems for Intel and the like) for example is using for their website.

Just wondering: why writing all the different filters again, when we already wrote them for AbiWord? AbiWord can even be used as a file format convertion daemon (non-graphical of course): dump a document to it and let AbiWord convert it to anything you like; plain text for example, so Lucene can index it! That would give you instant indexing/search capabilities for document formats such as Microsoft Word, Corel WordPerfect (complaints to me when a documents does not convert :), superb RTF support (only MS Word itself has better support for RTF, really), and a lot more. How nice would that be?

That reminds me: I should finally put an initial AbiWord SDK in our CVS and on our homepage, so everyone can see how easy it in fact is! Maybe tonight...

Abiword

Finally got AbiWord v2.0.7 officially released, after building some RPMs, writing the release notes and creating the changelog, yay :) Maybe I should do a v2.1.3 release soonish as well...

GUADEC 5

Still not sure if I'm going this year; my g/f can't come since she has to work, and my financial situation is about as good as the stability of the Mozilla M8 release. Maybe I should just go: I used Mozzie M8 as well :)

5 May 2004 (updated 5 May 2004 at 01:29 UTC) »

We got abisource.com finally back online, with hub putting a lot of time in restoring all our services. Rock! The University of Twente was kind enough to let us host all our services there, such as cvs, www and bugzilla. We only need to restore our mailing lists now...

Yesterday, me and my gf explored some parts of the Real World, and visited this artifical-tropical-environment-zoo-like place called Burgers' Zoo. The 'rain-forrest' is all indoors, with animals running freely around. It's pretty neat, especially considering that the most topical thing in the Netherlands would be a cow.

2 May 2004 (updated 3 May 2004 at 00:23 UTC) »

It seems that the nice Windows Viri running around these days are finally targetting the OSS movement as well. Yesterday our server hosting http://abisource.com, Abiword's developer cvs and Abiword's mailing lists has been taken offline: it generated more traffic than the network hosting the box could handle properly.

The maximum amount of traffic generated was around 40Mbit/sec. Quite a lot was most likely caused by the virus mails. Of course we block such mails, but you can't keep them from incoming...

About 11 months ago we added a mirror server to abisource.com for these kinds of events and is kindly hosted by the University of Twente (this is the University where I'm working on my thesis on the subject of Real-Time schedulers; in particular I'm studying/improving/implementing the EDFI and DMI scheduling algorithms in RT-Linux). By switching our services from our main server to this mirror, we hope to restore our services as soon as possible.

... and as you can see, I'm added to planet.gnome.org now, thanks Jeff!

25 Apr 2004 (updated 23 May 2004 at 19:12 UTC) »

OK, that took longer than expected, but I finally got Abiword v2.1.2 out the door. The reason that it took so long was that the ChangeLog was not ready yet. It was quite massive, and our not too descriptive commit messages hampered my progress quite a bit.

So, you might say, why not dump the raw messages in one big list and get on with it? I could do that, yes. However, I think it's very important to have a ChangeLog that mentions every single contribution that was made by our contributers. Every contribution can be very important, so it needs to get the proper attention. Let me explain.

I still clearly remember my 1st contribution. It was about 2 years ago. I tried Abiword every now and then, still being around version 0.7.12. For some reason, with every version I tried, I followed the same test pattern: Insert some text, Highlight a portion of it in the middle of the entered text, and hit View in webpage (or whatever it was called). Every release had the same bug: the highlight would continue up to the end of the line. This was always the point where I dumped Abi again, and got on with my life.

Until one day, I was bold (or bored) enough to download the source, grep through it looking for an HTML exporter thingy, and actually fix the bug. It took me about 2 full hours and it was a trivial 1-liner :). I started xchat and connected to #abiword at irc.gnome.org. Yay, people around. Heck, 3 even greeted me: hub, dom aka cinamod and Francis Franklin aka fjf. Neat. I greeted back. "Hi guys". "I've a fix for the Abiword HTML exporter". "dom: Great, could you send us a patch?". "I'd like to, but how to make a patch?" (yes, I didn't know a lot about OSS development back then). "dom: cvs diff -u <file>". "Ah, thanks". So I did, sent it to the mailing list. I hung around a bit for a reaction. "dom: Looks great, committed!". Didn't know what "committed" meant, but it looked like "your patch is accepted". After a "Thanks! Need to run!", I closed xchat, since I was already late for class.

A new version of AbiWord was released soon after the previous paragraph took place. IIRC, it was AbiWord 0.9.6. So I tried it. Usual testing sequence. Yay, my bug was fixed! No entry to be found in the ChangeLog though, so I was kinda dissapointed

However, this was the first time I gave something back to the OSS world, and it was accepted. By very kind and helpfull people as well. So I sayed around. Fixed some more trivial bugs that took me way too long to fix looking back. Currently, I'm active in nearly all area's of AbiWord's development. I add features, fix bugs, do releases, etc. And all because those 3 nice people greeted me, and made me feel welcome and my efforts appreciated.

So, to not to disappoint any contributers, I want the ChangeLog to be perfect. Maybe noone was in #abiword when some unknown contributer entered, and noone greeted him like hub, dom and fjf greeted me 2 years ago. At least he'll have the ChangeLog of the next release to look at, and see his name listed as a big "Thank You!" from the AbiWord developers. Hopefully, he'll stick around...

21 Apr 2004 (updated 21 Apr 2004 at 22:34 UTC) »

Working on getting Abiword 2.1.2 out the door at the moment. It's officially in feature freeze now, so it should get more stable by the day. I hope to push out some more development snapshots in a fairly timely manner. With a bit of luck, we can declare the unstable branch stable around July somewhere. Please give it a good banging before that time!

Trying to get AbiWord 2.0.6 officially released, but for a change, I want all binaries to be available at the moment of the release itself, and not at some random time thereafter. Currenly I'm only waiting for some FC1 GNOME rpms. Oh, and I should reupload the RedHat 9 ones, since for some stupid reason an epoch is set on the rpms that shipped with RH9... so, I decided to join the croud and throw an epoch in as well to smoothen the upgrade process.

... should be out real soon now.

6 Mar 2004 (updated 7 Mar 2004 at 00:56 UTC) »

In response to Havoc's blog on GNOME Office @ http://log.ometer.com/2004-03.html#6.

Havoc claims in his blog that GNOME Office was a huge missed opportunity. He's right about that one. The reason of it's failure is that GNOME Office has too long been this meta project, a project (if you can even call it that) with no vision. Recently, this attitude has been starting to change, and we are now discussing on the GNOME Office mailing list how to create the integrated, uniform GNOME office suite we all have been waiting for. We welcome input from everyone, so join the GNOME Office mailing list if you want to be part of the "new" office suite.

Now, Havoc says "Gnumeric and AbiWord are architected differently and use a different approach to portability, but why not suck it up for now and iterate them toward convergence over time".

Not sure if I understand the phrase "to suck it up" correctly as a native dutch speaker, but it seems to imply that we (the AbiWord and Gnumeric teams) are arguing over what would be the best approach instead of trying to work together an actualy produce something coherent. This is certainly not the case, and although I think Abi's approach to being a crossplatform application is the Right Thing(tm) to do, I do not mind at all that Gnumeric chooses another way of achieving it (ie. by using the GTK+ port for win32). No, instead, we *are* trying to integrate the applications better, for example by making copy/paste work between applications: try to make a table in AbiWord, select and copy it, and paste it in Gnumeric. It works. Of course, lots and lots of work needs to be done in this area, and progress is slow. Furthermore, Jodi started to create libgnomeoffice, and library which will allow us more easily to share code between the GNOME Office components.

The reason that progress on the interoperability front is slow is, at least in the AbiWord camp, mainly due to lack of resources. With an active AbiWord hacker team of about 7 people, yes that's seven, and some other more or less occasional helpers, we have to maintain a stable branch, develop new features in the development branch, maintain the webpage, do releases, help users, run through bugzilla, write documentation (which we actually don't have very much time for), etc.

This also explains why we don't make the "hard but worthwhile decision [...] to use the OO.org file formats" yet. Currently, we use the .abw file format as our native format. We are not against the OOo.org file formats per se, but adopting it would cost time, and lots of it. Time we don't have. This is mainly due to the fact that the .abw file format is almost a direct dump of our Piece Table, the mechanism used to hold the document internally. OOo.org's file format does not directly map onto our PieceTable, so importing and exporting it would take time to get 100% right. 99% is not good enough for our users. Another, slightly less important reason, not to immediately adopt OOo.org's file format is that it would make it a lot harder to do new, inovative things that do not map onto the OOo.org file format. We can now just "invent" the things we need for our file format, and get on with it. A process that takes about 10 minutes. We do have some initial work done on an OOo.org importer/exporter plugin, but developement is almost non-existant right now. If any hacker out their is interested into improving it, please join the abiword-dev mailing list, and we'll get you going. Work in this area would be appreciated by a lot of people.

As it is now, Havoc just has to wait a bit to "see the hard decisions and bold roadmap directions" be made. The (realy not so) hard decisions will be made, or have already been made. There are no arguments or flamewars between the different development teams that would hamper the progress we would all like to see.

A lot of good stuff happened lately. First of all, I released libwpd 0.7.0, which is a great improvement over the previous version. Being able to import WP5.x type documents seems to be a great hit :-) The plugin for AbiWord is already updated, as is wpd2sxw. Progress on the "WriterFilter", the plugin for OpenOffice, is progressing slowly as well, and a few success stories have already been reported on *nix.

On a very related front, I will have some very interesting things to report soonish I hope. Stay tuned if you are interested in everything that has to do with converting WP documents in *all their glory*.

Furthermore, I'm in the progress of releasing AbiWord 2.0.3, which will solves quite a few nagging bugs that made AbiWord quite unusable for a lot of people, such as Copy&Paste errors, and frequent crashes related to FootNotes. Hope this release makes everyone happy :-)

1 older entry...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!