A bientôt, là-bas ...]]>http://www.3pblog.net/index.php?entry=entry070512-1551092007-05-12T00:00:00Z2007-05-12T00:00:00ZTechnology topics of the blog have movedProcess-one.
The new content is available from my Process-one Blog Page.
See you there !]]>http://www.3pblog.net/index.php?entry=entry061115-1046262006-11-15T00:00:00Z2006-11-15T00:00:00ZHow to get started with wxErlangwxErlang: How to compile wxErlang on Linux.
wxErlang is an interesting initiative, as it could end up providing a very nice multi-platform graphical toolkit for Erlang. wxErlang is based on wxWindows.
The project has not been updated for a while and it would be very nice to see new contributions to it.
As an incentive, we have added the project to Process-one Erlang forge.
Ideally, this is the kind of graphical toolkit that could replace GS (based on TCL/TK) in the standard Erlang distribution. I think a new graphical framework can only take off if integrated in Erlang/OTP. This is the only way to make the developer sure that their graphical program will be usable on any Erlang installation.
From an Erlang developer point of view, the wxErlang code is also a very nice example on how to interface Erlang with C++ libraries.]]>http://www.3pblog.net/index.php?entry=entry061022-1046022006-10-22T00:00:00Z2006-10-22T00:00:00ZMetaforge Erlang at Process-one This tool was called Metafrog and was hosted on Erlang-projects website.
The work was huge and I had not enough time to make Metafrog stable enough, so I stopped the project. The SVN support was also missing and it needed more work to make it usefull with the new types of repositories.
Recently, I have discovered that a company called Cenqua has made exactly what I had in mind, in a product called Fisheye (Actually they do not yet do everything I had in mind but are working on Crucible to complete the features set with code review, comments and annotations).
This is a tool we needed for the various internal projects we are working on.
As a follow-up to the Metafrog project, we decided to offer to host Erlang Open Source project on our forge platform, as a service for the Erlang community. The hosting does not imply changing the projects approach or method. The repository can be kept where he is, but we simply offer repository browsing and analysis.
Please, drop me a mail if you know other Erlang projects that should be hosted on our forge.
]]>http://www.3pblog.net/index.php?entry=entry061003-1600412006-10-03T00:00:00Z2006-10-03T00:00:00Zejabberd and the Google Summer of Code 2006 The original idea was to create a proof of concept of a Wiki providing a realtime updates and propagation of the changes to the users currently viewing the page. With this kind of application, you end up with a Wiki that can act as a sort of collaborative whiteboard.
The prototype uses ejabberd publish and subscribe integrated module. Each Wiki page is mapped to an ejabberd pubsub node. ejabberd pubsub node being hierarchical, this maps easily to a website organisation. Each user currently viewing the page is subscribed to the corresponding Wiki node and received the changes to the page as they are made.
The project has been proposed and written by Grzegorz Grasza.
A demo site is available on the Realtime wiki demo page. Playing a bit with it can help understanding the underlying concepts.
The "Realtime wiki" project report is available from Jabber.org wiki (sorry, not realtime yet ;-).
We hope to have time in the future to contribute to this code, by making it more userfriendly for the end user. However, it is still a good validation of the idea.
Your feedback and ideas are of course welcome.]]>http://www.3pblog.net/index.php?entry=entry061003-1427292006-10-03T00:00:00Z2006-10-03T00:00:00Zejabberd 1.1.2La nouvelle version est disponible sur le site de Process-one: ejabberd 1.1.2, avec la lists complète des améliorations
J'en profite pour signaler un petit tutoriel intéressant sur l'installation et la configuration d'ejabberd sur Ubuntu.]]>http://www.3pblog.net/index.php?entry=entry061002-1652522006-10-02T00:00:00Z2006-10-02T00:00:00Zejabberd 1.1.2 has been releasedThis is a major version containing many improvements.
It can be downloaded as usual from Process-one web site.
The release notes as well as the complete changelog are also available.]]>http://www.3pblog.net/index.php?entry=entry060927-2357262006-09-27T00:00:00Z2006-09-27T00:00:00ZRuntimes for web applications It is a fact that the Web 2.0 is powered by runtimes on the client-side. Whatever the technology you want to use for Web 2.0 applications, you need a runtime. For pure Javascript based 2.0 application, you need a suitable browser version. This is the same for Flash applications, Java Applet-based application and XUL-based applications.
All Web 2.0 applications being build on runtimes, my point was to show that developers should feel free to consider all available frameworks for the task at hand. The developers cannot make the assumption that the correct runtime will be available on the client using the framework unfortunatly. There is a great chance that Flash 7 will be there, but in some cases, it will not be on the client desktop machine.
If you cannot be sure anyway the runtime will be there, then your duty is to select the best tool for the task at end, to provide the user with the best application, that will make it worth installing the runtime.
I see two main criteria that can make a difference when choosing a framework / runtime target:
- The ease of use of the end user.
- The ease of development for the developers.
It happens that those two criteria really depends on what kind of application you want to develop:
- If you want to build a multimedia application, Flash is probably the most suited. Sounds and drawing ability are the main Flash strengths. The development model is also quite suited to this kind of application. The Flash development model however makes it difficult today to development desktop like applications as there is no high-level widget available in the standard Flash client (Flex does exist but requires a special stack on the server-side. OpenLazlo is promising but still young in my opinion).
- If you want to extend your website with some interactive features, a pure web 2.0 approach might be a good solution, as it is often easier to integrate with your own site. The technology and expertise are already available in your development team for have this kind of development so it might be the wisest choice.
- If you want to build a desktop like application, going the XUL way sounds a good idea. You have access to true client-side widget and the programming model is quite easy. This is the path taken by AjaxWrite and the result is very promising. The Firefox extension model make it possible to use low level API, like in the Wengo Firefox extension for Voice Over IP.
If your creative enough, you can even mix several approaches together: A XUL client can embed Flash applets. You can thus build a desktop like web applications, using the Flash advanced multimedia feature.
Again, this is a general approach that depends on what your actual needs are. As often in computing, there is no silver bullet, only tools at your hand.
]]>http://www.3pblog.net/index.php?entry=entry060925-1849382006-09-25T00:00:00Z2006-09-25T00:00:00ZFlash, Jabber and Web 2.0Flash fat belly.
Actionscript 3 surely fixes the Flash hack necessary to connect to Jabber / XMPP servers: To ease parsing Flash clients were using null characters to separate XMPP packets. This was an ugly hack and it is nice to see it disappear.
However, I do not agree with the statement saying that Flash solves Web 2.0 compatibility problems:
«XIFF and Flash together make a very strong contender for building rich client XMPP enabled application without suffering this dreaded message one see on so many web 2.0 mashups "your browser is not supported, you must use [insert browser name here] instead" when trying to bring up the chat window.»
Flash only displaces the problem. Multiplatform support is lacking: Linux users can only get a Flash 7 player wheres most Flash applications requires Flash 8 or Flash 9. The other problem is that you still have to upgrade your player to be able to use the latest applications that only support the latest version of the player.
Flash on those aspects is not at all the silver bullet and things might get even more complicated with the next version of Microsoft and Apple Operating Systems coming out: Windows Vista will push its own "Web2.0" XAML framework and Apple will introduces web oriented Widgets.
Xul might also be an interesting contender, being truly multiplatform.
I think the future of rich web frameworks is still very fuzzy.]]>http://www.3pblog.net/index.php?entry=entry060924-1503322006-09-24T00:00:00Z2006-09-24T00:00:00Zejabberd 1.1.2 betaejabberd release.
This release can be considered a major improvement as it improves the robustness and stability of the platform.
This version promotes many changes since ejabberd 1.1.1. Improvements include:
- Major improvements on the LDAP module. It is now more flexible and should work in a much larger number of cases.
- Roster management improvements to make it more reliable (especially in cases where users are on different servers).
- Improved robustness: It is now possible to limit the number of opened connections for a single user.
- Database support: Microsoft SQL Server is now officially supported (ODBC mode).
- Improved MUC component.
- Shared rosters are now more reliable.
- Anonymous login bugfixes.
- Some protocol compliance small bugs have been fixed.
- The command line tool ejabberdctl has been improved.
- The build chain has been improved, including MacOSX support.
- A Czech translation has been added.
This version does not however introduce lots of new features, as this is what is being actively prepared on the next release, we hope to be able to publish for the end of the year.