On Fri, 2004-03-26 at 16:33, Ian McKellar wrote: > On Fri, 2004-03-26 at 15:11, Shahms King wrote: > > Yes, having the GNOME core implemented in multiple languages is just > > begging for the Sawfish problem all over again... With GNOME 2.0 right > > around the corner Sawfish had yet to be ported from Gtk+ 1.2 because it > > was largely written in Scheme and the gtk+ bindings hadn't been ported > > yet. > > Its quite different. rep was invented by jsh and is unused outside of > sawfish and jade (the editor John wrote - that only he uses). Python is > widely used, widely known, widely supported. > > Ian Yes, I love and use both Python and the GNOME Python bindings on a daily basis and would love to see them both become "Platform Languages" as it were, but my point still stands. Keeping the number of GNOME Official Platform Languages (tm) helps ensure continued maintainability of that platform should any particular developer stop his work on it. I agree that something is needed, I don't see any particular reason to stop using C/C++ as the statically typed language of choice for the immediate future, but having a popular dynamically typed (and garbage collected) language as part of the core platform should definitely be considered. In terms of the Java/C# debate, the way things stand now with ikvm and mono, any components written for the JVM can be used by the CLR crowd with relative ease while the opposite may not be the case. This leads to an interesting problem, given the popularity of the JVM and ikvm's ability to translate it for the CLR, it's logical that the JVM and Java should probably be adopted. However, the one-way flow of effort implies a wider variety of options appearing on the CLR side (Dashboard, for example). If someone were to write a CLR -> JVM translator the point largely becomes moot (except as relates to the "Maintainer Problem"). Whether to ship JVM or CLR binaries is then up to the distribution. Sun and IBM can ship the JVM compiled binaries, Novell the CLR. Admittedly, both cases depend on the completeness of the underlying Java/.NET implementation. -- --Shahms
Attachment:
signature.asc
Description: This is a digitally signed message part