Mailpost:earlystory

Originally from:

http://lists.ximian.com/archives/public/mono-list/2003-October/016345.html

Miguel de Icaza miguel@ximian.com
13 Oct 2003 12:48:58 -0400

Hello,

> Forgive the ignorance but did mono start their implementation before
> dotGnu of .NET? I am only curious. 

Well, we started working on the system about the same time.

On the Mono side, the events were approximately like this:

As soon as the .NET documents came out in December 2000, I got really
interested in the technology, and started where everyone starts: at the
byte code interpreter, but I faced a problem: there was no
specification for the metadata though.

The last modification to the early VM sources was done on January 22
2001, around that time I started posting to the .NET mailing lists
asking for the missing information on the metadata file format.

While I waited for this, I started developing the C# compiler as an
exercise in C# first a tokenizer, then porting jay to write a parser. 

Rhys contacted me about this time, he had been reverse engineering the
file format and had some early code to load the files.  It was an
interesting effort, and there was some early cooperation between a group
of three people: Rhys, Saurik and myself.   

About this time Sam Ruby was pushing at the ECMA committee to get the
binary file format published, something that was not part of the
original agenda.   I do not know how things developed, but by April 2001
ECMA had published the file format.

At this point the C# compiler was able to parse itself, and I demoed
this at the Guadec conference to a few folks.   Also at this point we
were able to do a feasibility study on the completeness of the
documentation published to build an open source technology.

Our feasibility study included building a metadata reader, which caused
much pain: Saurik had already done one, and felt his code was not being
used and Rhys had his own, which I did not personally like (for simple
reasons: it did not follow the Linux/Gnumeric coding style I used).  At
this early stage there was very little in all of these projects. 

Since December 2000, we had been amazed by the .NET Framework, and when
we discussed internally at Ximian its benefits, what we initially did
was to staff the "Labs" team to work on CORBA, SOAP and Perl teams to
work on the Gnome binding infrastructure (remember: our motivation at
this point was the vision of writing APIs once, and using them in every
language).

The Labsl team effort's eventually resulted in work in Bonobo-conf,
ORBit2, bonobo-activation, Soup, and the Perl/Gtk bindings.  The
intention was to build tools to improve our productivity: create more
applications in less time, bring more abstractions and standards to the
desktop and reduce our time and cost of development.

The team products were positive, but still fell short from everything
the .NET framework could do. 

But when we were done with our study, it was clear that it was possible
to build this technology, which we consider key to the future of Gnome
and Linux on the desktop.  Remember: we were developing the largest
from-scratch desktop application at this point, Evolution (Mozilla and
OpenOffice are open source, but were originally proprietary products,
later open sourced).

So we had some experience on building open source projects, and we had
a relatively important source of pain that needed to be addressed. 

Nat Friedman was highly supportive at this point of moving our efforts
into something that would have a larger impact, and once we got the new
management at Ximian (David Patrick joined us as CEO), most of the the
developers from Ximian Labs were turned into what became the Mono
team.  The objective of the team, just like it was before was to build
tools to increase programmer productivity.

We remained quiet, as we moved the teams over from their existing
projects to the Mono effort, they were winding down on their existing
projects, and only a couple remained behind: Alex Graveley (building
Soup) and Michael Meeks (working on Bonobo and ORBit).   The rest,
Dietmar, Paolo, Dick and myself started work on Mono.

Ravi will join us later to assist in the C# compiler development.

It is obvious that a small team like this can not build a full .NET
replacement, so we planned to launch this as an open source project,
under the direction of Jon Perr in marketing that helped us get the news
out that we were going to build this project.

We planned the announcement to come by July 19th 2001, so we could
announce this at the O'Reilly conference, as Tim O'Reilly had been very
supportive of this effort, and had offered his help since the early
stages, when it was still a very young idea.  When we announced the
project launch we had our team in place, and we were shipping our
metadata framework and our C# compiler as well as a few initial classes

So officially the Mono project was launched on that date, but it had
been brewing for a very long time.

Who came first is not an important question to me, because Mono to me is
a means to an end: a technology to help Linux succeed on the desktop. 
Of course, it has taken a life on its own, because Mono is not what
Ximian/Novell chooses it to be, it is the result of the contributions
and opinions of its contributors and users.  So Mono has grown larger
and better thanks to that. 

Hope that answers your question,
Miguel.