[Mono-list] Re: Mono 0.25 and MonoDoc 0.5 released.
Miguel de Icaza
miguel@ximian.com
28 Jun 2003 23:23:35 -0400
Hello,
> The progress on Mono is amazing.
>
> 15 rpms for three redhat distros, 5 for Mandrake, 9 debian packages --
> 29 in all. One Windows setup wizard with a relocatable install that
> runs on Windows releases pre-dating all of them (and most of Linux).
>
> <wish>
> I hope someday Linux can figure out how to do this. Ximian, please
> help us developers find a way to independently build and distribute
> one binary, which is distributed in a single installer, and which
> yeilds a relocatable package, that can be installed on a large
> percentage of Linux machines, that remains compatible for a couple
> years.
> </wish>
This is the very same wish I have.
Today we aim for "best integration with existing distribution", and
that is why we ship custom-built packages for each distribution: system
dependencies are acknowledged and we link with the particular versions
of each Linux system.
I have to say, that it is a problem. For instance, as much as we would
love at Ximian to support even more distributions, we just do not have
the man power to build, test and provide support for the ones we have.
Each new distribution requires custom work.
Maybe .NET and Mono will at least simplify this pain by having "xcopy"
deployment. But it wont solve all the dependency issues until all
distirbutions agree on package naming conventions and on a basic core
(it is still a valuable thing that dependencies exist).
It seems to me that Linux still needs to mature on this area. Today
most distrbution players focus on delivering the best value for their
users, but they are all competing for market share, there is little
incentive to standarize on a core distribution.
Also, not everyone understands the value of backwards compatibility. We
should start pushing people to be more aware of this. This idea took a
couple of years to "permeate" through gnome. Gnome has been an
interesting vessel for various ideas: usability, accessibility, ease of
use, binary compatibility and enterprise features. Each one of these
ideas started to permeate, and it did take a few years for the ideas to
mature into things that people were aware about on their day to day
hacking.
But we should push for this agenda in even more areas in the free
software world, or we will continue to punish the ISVs.
Miguel