Older blog entries for mpesenti (starting at number 19)

7 Apr 2004 (updated 7 Apr 2004 at 11:33 UTC) »

My immediate reaction to Brendan article is mixed. Braindump follows.


- Mozilla foundation seem to be considering Linux like a primary target ("Let those enterprises migrate to Linux if it makes sense, or defer the hefty Longhorn upgrade tax by sticking with downrev Windows for as many years as makes sense."). This is a fundamental prerequisite if we want to build the future linux platform together.

- It seem to anticipate a different approach to multiplatform where native, available technologies are reused (while filling gaps on some platforms)


- Migration to the future platform need to be gradual. For reasons that I dont fully understand mozilla.org is advocating the substitution of current linux desktop applications with XUL based counterparts. The merge need to happen on paritary basis, gnome libraries or mozilla libraries are an equally good way to write apps until Gnomzilla platform is available. I can see new Evolution features being written using the Gnomzilla platform (Mono + XUL for example) and mozilla rendering engine integrated in the old code using gtk wrapper (gtkmozembed). I cant imagine adopting Thunderbird though. The prioritary goal is to have a mail client well integrated with linux Desktop. My feeling is that what you can really share cross platform are just the engines, the UI needs to be special cased to be really integrated. But, even ignoring this, it doesnt make sense to throw away all users benefits of current solutions to adopt immediately and completely the new platform. A gradual migration path is necessary ...

- I fear the complexity of the merge is heavily understimated. Given the current status of mozilla SDK I'd be positively surprised if by "2nd half of this year" we would have something that can easily work outside mozilla tree, based on stable API etc. Let alone how much time really merging it with GNOME technologies will take. Because of this reason gradual migration is even more important.

Anyway I think it's crucial that the two communities start to communicate and cooperate. (Got my first epiphany patch by a mozilla hacker, yay !)

I feel like we are at an important time for the linux desktop future. I really hope the result will be reduction of the fragmentation rather than increasing.

BTW the gtkmozembed GRE work is progressing thanks to the feedback, the help and more importantly the patience of people like bsmedberg, blizzard, darin, biesi. But yeah yeah this isnt really exciting anymore, everyone wants Gnomzilla now, I'll see what I can do :P

31 Mar 2004 (updated 31 Mar 2004 at 10:51 UTC) »

I have got a basic version of gtkmozembed working with the GRE. I had to strip out some features because I cannot use mozilla internal api. I think using the SDK is a better approach then having an interface in the middle to work around the problems with the SDK itself. The interface in the middle solution is a bit like claim mozilla failure to provide an usable SDK :) The API problems seem to be solvable to me, I'm more worried about mixing glib and mozilla api for the strings, especially because memory allocation is not compatible. Maybe the minimal embed string api should have a way to do unicode conversions (utf8 <-> utf16). Oh well let's see if there is interest in this approach at least.

Now that I have figured out the GRE thing I'll be back working on multi screen. Mark suggested me a way to test with software only. Rocks !

Xnest -ac -scrns 2 :1

Today I have to meet with my thesis co-relator. Hope everything goes well. Wish me luck :)

30 Mar 2004 (updated 30 Mar 2004 at 11:14 UTC) »

We've finally been able to publish 1.4 plan: abstract and concrete, very much a work in progress.

My personal fight with mozilla SDK continues, well at least we are progressing.

Did some basic work on multi screen, my problem is that I cant really test it. Synap promised to help though. Anyway is there a way to do test without hardware ? Alternatively, could someone send me the required hardware :P

Finally a few day out of civil service. Staying 6 hours a day with old people is ... well ... stressing ;) I went to skying yesterday and this morning, I got some photos but I've been too lazy to put them online so far.

28 Mar 2004 (updated 28 Mar 2004 at 10:06 UTC) »
Open source contest 2004 is a very admirable effort to reward Italian Open source developers. The decision to involve also school projects is interesting, it would be great to improve connections and cooperation between the acamedic world and free software. Obviously I subscribed Epiphany :)

We have a plan for Epiphany 1.4. Unfortunately because of post hacked widget problems we have not been able to publish it yet.

I'm trying to make gtkmozembed an interface to GRE. That way I think it will become a much more viable candidate to be used in other desktop apps. Immediate advantages are the ability to use it without ugly wrapper scripts with the mozilla libraries directory hardcoded and to not depend anymore on the whole mozilla browser installation. In the process I found an interesting bug in the SDK. Well the request for blocking 1.7 sounds promising. Thanks bsmedberg ! (I wonder if mozilla guys reads planet gnome).

Mark cleared my confusion on multihead. Next week I should be able to add multi screen support to Epiphany.

Seth posted a patch to add non local uri support to epiphany file pickers. Unfortunately mozilla support for gnome-vfs uris is still not complete and we had to delay it post GNOME 2.6.0. For 2.8 I think the way to go is pretty clear: complete mozilla gnome-vfs support and allow non local uri for almost all file pickers. On the short time (2.6.x) I'm not sure what's the best solution though. Mozilla 1.7 can already open all gnome-vfs protocols but not save them. The most conservative approach would be to enable local uri for open mode file pickers only if compiled with Mozilla 1.7. Otherwise we could make sure the operation doesnt fail silently when saving/opening a not supported protocol and show a warning to the user. Is the ability to save to webdav and sftp worth the confusion of showing smb mounts in the filepicker and then failing to save in them ?

I love the new bugzilla. Thanks to everyone that worked on it :)

19 Mar 2004 (updated 19 Mar 2004 at 23:55 UTC) »
jdub They must be confusing me with someone else ;) Actually noda suggest that I'm famous as "that idiot that forked galeon" :)

As everyone else I've been following the languages discussion but I still havent build a definite opinion. I think it would be important to understand how much the ability to run managed code from the C/C++ code will be useful in practice. Anyway I hope to see a strong effort to find a common solution, fragmentation is a real and immediate risk.

I strongly feel we should keep very clear the distinction between platform and language. Using different platforms to build the desktop would have a strongly negative impact on usability (because of consistency but even more importantly because of integration) and code sanity. Havoc is addressing it very nicely in his document. When I hear "write once, run everywhere" I start to fear.

17 Mar 2004 (updated 17 Mar 2004 at 08:14 UTC) »

I'm really dumb sometimes. I released Epiphany 1.2 thinking that this Monday tarball due was for proper 2.6 and not for pre-release. Oh well ;)

Epiphany users looking for a js console should try out epiphany-plugins. There is also an html error checker and a very nice popup blocker (View->Show/Hide popups remembered per site and a notification icon in the statusbar, that's all).

16 Mar 2004 (updated 16 Mar 2004 at 01:25 UTC) »

I really hate when someone define mindless users that doesnt want, need or care to learn unix filesystem or shells. Maybe it's just because I'm that sort of person myself ...

I was reading Seth piece on usability tests and focusing on a somewhat marginal point. "Producing a good design requires more art than method". I agree that creativity has a very strong role in interaction design. Though I was wondering how this art could be better communicated and teached in the free software context. While theory can be easily transmitted by books, essays, or blogs entry, artistic skills are problematic. Like crafts they could probably be better learned by working with the "artist" (the designer). Free software is probably the most succesfull example of collaborative work nowadays, which make me think we have a great opportunity but we are mostly missing it. Why ?

Oh btw. Epiphany 1.2 is out ... let's hope everything is ok.

I've been doing some work to add a gtkmozembed backend to yelp. There are a few issue but they seem to be solvable. What I dont really what to do is to add yet another wrapper script to setup LD_LIBRARY_PATH/MOZILLA_FIVE_HOME. Mozilla embedding really need to be fixed. I wonder if blizzard would mind some help on the GRE port while he is busy with other things :)

It's snowing crazily here. Never seen so much snow in my country before.

What happend in Spain is terrible. I have really no words.

I had a one month unexpected delay with my thesis work. Though finally things are shaping up again and I'm not far from finishing. I hope everything is ok this time.

We had strong bug submission attacks in the last week but we managed to resist. It's fun to be able to start working on bugs as they are submitted because of the challenge feeling. Actually Christian usually fix bugs 15 minutes before they are submitted, so all that I need to do is close them :) I think 1.2 quality will be good but we will have to wait Beta 1 release to have a confirmation.

I've been sucky lately with my private mail. If you wait for a reply do not give up yet, next week I should be able to go through my queue.

24 Feb 2004 (updated 24 Feb 2004 at 11:27 UTC) »

I've got Where the action is from Amazon. Had a fast read and it looks cool, now to find time to read it for real. I'm also reading Norman's Emotional Design that my gf gave me, lots of interesting points there too. I must say I envy Edd to have listened Norman talk :)

I feel sad about the problems translators are having with string changes since I personally "contributed" with two strings (one of which was really really silly). Sorry guys.

Glad to see the new file selector design. I especially like the behavioral rationale. Gnome 2.6 file management is going to rock with this and spatial nautilus !

It have been snowing pretty heavily here in the last few days. Today I have no civil service work, so I plan to go out skying in the afternoon. Or maybe I should fix some bug ... nah ;)

10 older entries...

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!