Re: Feta font

[Juergen Reuter]

On Wed, 3 Dec 2003, Joel A. Erickson wrote:

> ...
> The idea is to create a simple web-based notation editing/proofing IDE for
> collaborative proofing work. It's basically a human-powered optical music
> recognition program. The output will be data, such as MusicXML and/or
> Lilypond.
> ...

In 1996, I started a project for collaborative versioning / configuration 
management over the world wide web, using standard browsers on the client 
side.  The first try was an almost pure HTML (with forms) based solution, 
using only two very small helper applications (~100 lines of perl) to 
gain access to the client's local disc.  This project more or less failed 
because of the various limitations of browsers at that time (e.g. tricky 
dirty and very slow handling to gain access to local disk; statelessness 
of the HTTP protocol, which is not session-oriented; therefore explicit 
session id management (without cookies) through URL encoding; browser's 
back/forward buttons disturbs tracking a session; ending a session by 
closing the browser without notifying the server -> need server-side 
session timeout -> may kick off slowly responding user...).  A later 
version based on Java applets could circumvent some of the drawbacks, but 
also suffered from various restrictions of browsers at that time.  A third 
version of this project finally still used HTTP (or, to be more precise, 
an early version of Delta-V, which is an extension of the WebDAV protocol, 
which is itself an extension of the HTTP protocol) as underlying protocol; 
however, this time I added collaborative functions to an existing 
stand-alone application for versioning, using existing HTML/HTTP 
libraries.

Of course, browsers have changed since then and many things are possible 
now by cookies, Flash, JavaScript, signed applets, ActiveX, etc. -- but at 
the loss of the original spirit of the concept that HTTP+HTML+browser 
gives you a platform-independent framework for the GUI and access to a 
remote server.  I think, if you are trying to implement a fully-fledged 
interactive, session-oriented application with a complex GUI, you will 
spend most of the time to use tricks of today's browser technology to 
circumvent the fundamental conceptional restrictions of browsers.  And 
most likely, this will result in something, that is not at all platform 
independent; that is not the idea of collaboration.

Rather than adapting a browser to collaborative music notation, I strongly 
suggest to adapt one of the existing music notation programs (e.g. one of 
those that Jan mentioned) to collaborative work over the web.  Why not 
writing a web module for one of these programs?  Otherwise, you will 
effectively start a complete music notation project of your own within the 
browser framework, which may take you decades of time (this is not a joke: 
LilyPond started more than 6 years ago and still has fundamental bugs 
(e.g. collision problems) and a long todo list!).

Greetings,
Juergen