On Tuesday 07 January 2003 21:20, John Gage wrote just slightly over the top in places:
> Brian Bray, the founder of the list, has already said that open > source must follow the taillights of proprietary code. I think we would all agree that in many areas OSS _has tended_ to follow the taillights, but I don't think anyone says it _must_. Heitzo makes a reasonable point, but ... > >As a programmer for some 30 years and with a lot > >of languages under my belt, I'm not interested in > >learning another language that doesn't move me > >forward in my career .... I have enough dead languages > >already in my brain. there are careers and careers, and, say, quadrupling the size of a universe of jobs in M may be easier than developing that much career opening in new stuff. Maybe. But that isn't the main point I'd like to put out for inspection. In the RAD and extreme programming talk there is mention of _refactoring_ and that seems a sensible way to do things. So, where a component of size from a small Perl script to a group of modules or the whole of Vista is doing its job, leave it. If M sinks slowly into the strata of the past, then if programmer-archeologists are becoming scarce, is the time to take a single module and refactor it into a different language. There is a risk of course that as you refactor it you become adept in the old language, and therefore prolong that part of your work, as you succumb to the temptation of fixing something rather than replacing it here and there, but I like it. The result of this for the user is that from time to time a new function apepars on their screen, but nothing goes away, and there are no big upheavals, just a steady upwelling. A feeling of a live system rather than earthquakes. Historically, M has worked well, and projectes constructed using M in health have I am told failed less often than others. I don't have figures, but I do remember a specific paper in the British J. of Hospital Computing some years ago. -- From one of the Linux desktops of Dr Adrian Midgley http://www.defoam.net/
