|
From: Nuno F. <na...@al...> - 2003-07-01 15:38:43
|
Hi Marco and Miguel, Here in SIG Lab we all apreciate your good work and interest for the project. Now i will write some considerations in portuguese. Desde já axo que a ideia de dividir o projecto em três partes é boa, na medida em que me parece aceitável separar o desenvolvimento da interface gráfica do desenvolvimento do parser. A 3ª parte é a junção das outras duas para funcionar em PDAs (penso ser esta a vossa ideia). No entanto, e dado que voçês estão a pensar usar Java2 no vosso desenvolvimento, quero alertar para o facto de que o objectivo deste projecto é mesmo as plataformas móveis, porque já existem vários visualizadores de SVG "livres" (por exemplo, o Batik da Apache) que suportam toda a especificação SVG e que usam as mais recente tecnologias Java mas não funcionam nas máquinas virtuais de Java para as plataformas móveis. Daí, o nosso interesse maior é que o projecto avance para o funcionamento em PDAs e é aí que nós precisámos mais da ajuda da comunidade científica mundial :) pois o nosso conhecimento nessa área ainda é limitado. Entendo perfeitamente quais a vossas intenções visto que o mais importante serão os algoritmos usados na programação, mas queria que não se desviassem muito do objectivo principal do projecto e que também contribuissem (sempre que possível) para esse objectivo, que é por a funcionar decentemente o visualizador em PDAs.. Continuação de bom trabalho. P.S.: Temos de marcar uma reunião para discutir ideias e soluções. Pode ser pra semana? (Esta pergunta é pra todos os envolvidos no projecto... :)) Nuno Faria ----- Original Message ----- From: <ma...@ac...> To: <ope...@li...> Sent: Saturday, June 28, 2003 2:07 AM Subject: [Opensvgviewer-general] Cleaning done. > After 96 man-hours, the cleaning is done. > Results: line count down to 3333 from 13030. Speed increased, memory footprint decreased. > > The development should now proceed in two fronts: reimplementation of the GUI and reimplementation of the svg parsing. Miguel will work mostly in the GUI, and Marco in the parsing. The other developers are, of course, free to choose what they want to do. > > The reimplementation of the GUI will proceed in the current CVS tree (Viewer/). The reimplementation of the svg parsing and DOM construction will be started in a new tree (gvsj/). Both these trees will be coded for, and using, the most recent technologies. That means Java 2, including the use of Java 2D and the most recent Java libraries. The objective is to do a complete implementation of the Mobile SVG's Basic profile. These trees should be merged sometime in a not distant future. > > Of course we still want to run the viewer in pockets, which means the Java 1.1 platform. With that objective, a third tree (gvsj-fly/) will be created. The purpose of that tree is to change the code in the other trees to make it work in Java 1.1. Of course that doesn't mean we will have to change all the code from the other trees, but, instead, to create the necessary libraries to substitute the ones not present in the Java 1.1. It would be nice if someone steeped forward to manage this tree. Anyone interested? This task will not be very difficult: we, all, will be careful when choosing the code to use in the other trees to make this task of porting an easy one. > > For now, the development process should _not_ be closed. One option would be to have a maintainer for each tree that would accept patches from the other coders; no one else would have write access to the CVS tree. This is a good way to maintain a big project or one that is already in stable state. As we want to develop as fast as possible, and we sure are not big yet, the 'open access' is preferable. > > And that's it for the plans. We are really hopping for replies to this message. > > P.S.:Due to the change in the GUI, the attached file is needed to run the application. It must be copied to the directory '../icons' relative to the location of the code. That is, the directory containing the code should be sibling with the directory 'icons'. > > P.S.2: The numbers are a little misleading, and a little provocative to the original developers. Many lines came out, but we trashed some functionality in the process. To maintain the same functionality as the original code we would need, for sure, at least... lets say... 4444 lines of code. So, they are not as bad as they seem, looking just to the numbers. > > P.S.3: (Quem escreveu o P.S.2 foi o Miguel. Eu n?o tenho culpa. --Marco) > > Miguel Castro > Marco Monteiro > |