Re: Not snazzy (was: New Unicode Savvy Logo)

From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Sat May 31 2003 - 14:15:37 EDT

  • Next message: Karl Pentzlin: "Re: Fw: Unicode filename problems"

    From: "Marion Gunn" <mgunn@egt.ie>
    > Ar 17:51 +0200 2003/05/29, Philippe Verdy entre sur son clavier:
    > >I would prefer to say that Netscape 4.0 is dead, but Netscape 4.7x is not (I
    >
    > D'accord. (With the above I'd have to agree.)
    >
    > >see no reason why users should continue to use versions before 4.7, as the
    > >4.7 version fixed a lot of interoperability problems, including
    > >cross-platform compatibility with other Netscapes, plus many security
    > >fixes...)
    >
    > Yes.
    >
    > >Netscape 6+ is still too new with its new operating model, and lacks the level
    >
    > Again - after some scary experiences with 6+ - yes, I'd have to agree.
    >
    > >...
    > >However the recent versions of Netscape 7+ based on the new Mozilla Gecko
    > engine include a lot of performance enhancements in the JavaScript
    > >engine,
    >
    > Really? Most mail I get seems to be generated by MicroSoft Office slaves.

    I don't think we were speaking about email agents. For me the new Mozilla Gecko-based browser is great, but the mail agent is too crappy to be usable... Most of the mails I reaceive come from users of Outlook Express which is just fine for what it does (I don't speak about Outlook in the Office Suite, which is just a open hole to the system, and a huge resource drain).

    I'm not a slave of Office products (even if I use it because I already have a licence of it and I need something that can process the most complex Excel worksheets.) I use other word processors too. I will never buy Office XP or Office 2003 (Office 2000 is just fine for me). But my experience with OpenOffice were very deceptive (too much resource intensive)

    > >Only stable parts of the development are optimized, to avoid creating
    > >>unmaintainable source code.
    >
    > There you lose me, as I do not comprehend the above sentence - could you
    > rephrase it, perhaps?

    An application can contain a lot of performance improvements by including some stable components on which much work has been done (for example newer versions of the Xerces XML parser, and new internal data models for processing). Still there's a lot of unstable parts in the code that has knownbugs that still require a lot of work (notably the XSLT layout engine, CSS3 properties and their bindings to Javascript including security restrictions, and new integration libraries to support faster rendering such as support for DirectX on Windows within the layout engine).

    What is not fully optimized is the Javascript interpreter engine which was entirely rewritten to support other component integration models such as COM, Corba, Java, or synchronization and events with these components in addition to the legacy plugins model.

    Also in this area, the support of Unicode normalizations and transformations is quite slow and could be updated to support more recent versions of Unicode. Finally the built-in Java engine has not be tuned up specially for a good integration with the browser's usage of Java in a separate VM for its GUI interface, and all the works with skins in Mozilla was probably not a priority for Nescape 7. Nescape should have beter focused on demonstrating its capability of driing the Mozilla project to meet the indusry standards, without surcharging the browser with non critical components.

    Now that AOL/Time Warner has signed the final agreement with Microsoft who will give free licences of Internet Explorer for 10 years, will AOL continue to invest to maintain the Netscape browser application and helping the Mozilla project? I hope that Mozilla will be able to continue as the best reference platform for conformance to open standards such as W3C's. But here's now the risk that the W3C will be too much influenced by Microsoft's solutions.

    May be it's high time that the W3C adopts a more international vision of "open" standards, using less restrictive access rules for its decisive work groups.



    This archive was generated by hypermail 2.1.5 : Sat May 31 2003 - 14:58:54 EDT