xuland

Asa Dotzler, a person who loves mozilla and now a mozilla employee will visit Argentina next week, on July 2nd and 3rd. There are plans for at least a couple of events and all will be really important, I encourage you to assist.

Please, remember, this is the first time a mozilla person travels to this country, keep in mind that fact and you will enjoy it the event twice. To more information, and to register you, please, see the links below.

http://www.cafelug.org.ar/modules/eguide/event.php?eid=1
http://www.cafelug.org.ar/modules/eguide/event.php?eid=2

I want to thanks Asa for this beauty opportunity in Argentina, and to the organizers:
Javier Arce
Augusto Daniele
Daniel Coletti
Felipe Lerena
Nicolas (I will update with his last name soon )
Marcelo Poli
Maria Arce

Enjoy it, and see you there ...

Well, I am volunteering some time to mozpad. I hope that can be useful to somebody. Also, I want to do somethings...

I was working on a tutorial form some time, that I am going to finish as soon as possible, about writing a c++ xpcom module -another mysql wrapper (some day I will understand why we love to do wrappers to mysql ).

By the way, I am going to start a little program to make easier the development of xpcom modules, I am thinking on an interface builder tool. In my mind I see it like a little "uml designer view", then you press a 'make' button and it (1) creates the .idl and (2) runs the idl compiler).

Some thing I still do not know is how to run a bash/console aplicattion from an xpcom and get back the output, may be it will be necessary to develop an xpcom for that. To have a "console inside xul" would be enjoyable.

Today ( 2007-05-23 ) was the kick of meeting of mozpad over irc... It was amazing the speed of the talk, and for not English persons a little complex to follow and contribute in real time.
At some point why mozpad exist? what are the goals? If "It is a community of companies, organizations and individuals that are using Mozilla technologies to develop their products." May be a good answer can be ' to make life easier to' "companies, organizations and individuals that are using Mozilla technologies to develop their products".
And the big questions is... how? Well, we are doing stuff, then...

To do, we have to know.
As the talk evolves, appear that to provide documentation is not enough, currently there are some docs on mozdev and xulplannet.
At some point, we are playing with a big piece of code, that is complex, and hard to understand. In example, if you want to use mozilla and need to extend the widgets with a complex one (in the same level as the tree widget), that can be a difficult task, how does gecko work? what is the right way to extend it?.
A good way to learn is imitation, we can develop tutorials, or a kb, or a forum for support.

To do, we need tools (starting at the basic text editor).
But, if you have enough knowledge of that you are going to do, doing it is still complex, may be you have to write xbl's idl's, cpp's, makefiles, use a couple of macros and a lot of things.
Then tools can make you life easier, an interface builder, an text editor that highlight not only the c/idl text, but the macros, or the correct types (some time ago I got an error that took me one hour to find that the error was misspelling PRBool as PBBool.
Still, we can help with the mozilla build system, to make it easier (run the compile script of mozilla every time you change your code, can take a lot of time).

To do, we need some extra 'help'
Suppose that you are developing a mail client and want to preview the media attachments, you don't have an internal video viewer.
Suppose that you want to read or write compressed files, you cant (may be reading zip yes because that is used by firefox,xulrunner,...)
The point, we need to build libraries to extend the platform. I think that can be achieved by building wrappers of another open source initiatives. We can put those as add-ons for use of another developers.

To do, what more do we need to improve our doing?

This month I was thinking on write about another topic (probably you will read about it in a month ). But by those days there was a kind of storm in the mozilla community ( a community which, from my humble space, I try to form part).
First, a words on mozilla and Firefox. Firefox is a web browser famous by its merits on the "easy to use", the "secure" and the "powerful functionality and extensions" fields. It is built on top of severals technologies, that are so powerful and modular that the foundation has decided to make it more accessible to another developers and projects, and that effort and framework is known as XulRunner.
Someone said that because mozilla had a great technology platform, they have to invest more in it and left in second plane Firefox.
I agree in the real value of xulRunner, here in HS we use it and we only have nice words for it. But I agree that the mozilla suite in its days, and Firefox nowadays are the engine that bring that technology up to now.
We can not kill firefox, because it is valuable for us, and for the future of xulrunner.
I think that one of the points because I love open source is that if you want and can, you can put hands on and make a difference.
If we love Firefox and xulrunner, we only have to invest time and effort in doing both better. Telling to everyone that Firefox is amazing, helping some person to understand some part of the mozilla base code, writing tutorials and extensions.
We can help in doing mozilla better, not because they are in an error, but just as a way of saying 'Mozilla, this is my way of give you "thanks"'.

Just a note, remember, mozilla had a little amount of resources (I think there is about 50 employees), and with that in mind, it is amazing all the things they are doing and getting success on.

The video: http://factoryjoe.com/blog/2007/05/10/thoughts-on-mozilla/
Some nice answers:
http://weblogs.mozillazine.org/mitchell/archives/2007/05/application_vs_platform_focus.html
http://www.allpeers.com/blog/2007/05/13/only-the-paranoid-survive/
http://www.allpeers.com/blog/2007/05/11/mozpad-a-mozilla-platform-user-group/

When you are developing a piece of software, you want to do a beauty work. We want to develop something meaningful, something that can be useful. When we started building prototypes, we felt the goods of that code, and the harmony of that architecture start to appear.

When the time has come, we are ready with a nice soft, (maybe a library, an executable, a whole application, that does not matter). And in every adjust we made it, in our own mind we think in all the story behind that code, in which was the first attempt, the corrections we had to make, the things that did not work in some environment or configurations...

If you team are planning to release it as open source, then (does not matter how fine it does its work) you probably have done a great mistake, the lake of documentation. Yes, you can thing that releasing the code is all that a person needs for understand and extend or adapt your code, that is right. But what happens if you release documentation, really good documentation? Well, the learning curve of how you have thinked the soft, how you have solve platform and configuration issues, what is more or less the scope of the soft, and finally how to understand the class and file distribution of code and roles, can be more soft to take, I mean, will be more easy to learn that things from clear documentation than from your code (still we can learn that from the code, that is not the point, the point is the developer experience at working with your project and the developer time invested in learnt it).

In a close software environment, it is the same, new developer will have to deal with that code, or that piece will be created, maintained and updated by different teams. And no always you will have the chance of use a developer's time to teach another developer the internals of a piece.

Then in both cases, it is important to produce clear, mature and abarcative documentation about all the relevant parts of a project. For good of everyone, in the company, or in the open source community.

There is a lot of possibles ways to build mature software, but one is special for me. It is very powerful, and it is clearly a common tendency right now.

It is the art of provide a Aom or a Dom, [application or document object model]. The point is to provide an easy way to allow some kind of control by the user, to increase her/his power and to permit a developer to integrate and/or communicate with the piece of software in a more soft way.

It is wonderful, and if the a/dom is well done, it is like craftsman. But there is a problem, sometimes it is no very well designed, or there is a crypt way of do a very simple o common task, or in a innatural way of thinking how to use it.

It is really important to understand how the user of the dom will really see the problem to solve, and to provide that kind of xom.

Some times, there is an internal dom, available from the plugin system or to use by the application (in that case it is only useful for the team that build it). And the external, that allow you to control the program with the same power like using mouse and keyboard -yes, that can be dangerous from a security point of view.
Some examples of real word xoms:

- Gaim, by this time, Gaim (a popular multi protocol instant messenger ) it is separating the core functionality (libgaim) that deals with the network and protocol stuff, from the user interface, that means that you can provide in your application im support only reusing that powerful library, amazing, do you?
- Firefox, well, actually all xulrunner applications, but let me focus in Firefox. This brouwer allow to the plugins to access the core components and technologies that uses, that means that you can touch everything . And using xulrunner you can not only have a xom, you got a complete application framework ready to use.
- Blender. A 3d suite for animations, it exposes itself as a series of python modules to allow scripting inside the program, you can do everything and more, it is so magical that you would want to have the same set of modules to use in the published game (when you publish you work not like a movie, but like a game you can use a runtime subset from within the python scripts of your game).
- Open Office. It is beauty to integrate the uno framework into you project, it has a powerful framework to deal with office documents, and it is not so complex to understand.
- I know not so much about the internal of Gnome or Kde, but in the last one, it is very known that the kpart-technology allows integration and reuse of components be speaking the same language.
- Gimp, the best gnu image manipulation program , allows you to write script in scheme, and they can ve really powerful, but Gimp doesn’t have a ‘libgimp’ to be called from another program and use the power of gimp.

For me the lesson it is that: if the program can do a nice work, good. But if it can be ‘public’ in some way, called by command line as first attempt, then add dom and script support, well you develop will be a true inestimable tool.

The last comment about it, some times a good library with no user interface can be as valuable as a ui-program to the market. Why? Because the ui it is easy to build, but manage a complex file format, or a complex network protocol, it is another story, at some point, in some cases, they solve external problems of our soft, problems that need to be managed successfully, and that are important by they own, but that are not ‘the problem’ for us.

Anyway, script-ability (aom or dom) internal (for plugings) or external (to use for another program) it is a desire feature that we can enjoy in good software.