Em 21-02-2011 16:07, csrabak@bol.com.br escreveu:
Geert,
I'll use my consultant's hat to contribute on this. My company lives from selling advice on these matters.
Notice that in the table of the referred link: http://www.theaustralian.com.au/australian-it/legacy-languages-prove-lucract... which you point to, Cobol and Smalltalk share the same status of "Retire Now".
So the Channel Register interpretation of the reports has not been that sharp at all. Also, the company issuing the report is an Aussie centered firm, as the problems we're facing in the companies around the world (in the mainframe) is more what to do with the Natural/Adabas platform, where Natural is a higher level language than Cobol but has a high impedance for conversion to Java (w/o opening the can of worms of running Java in mainframes).
The FUD about Smalltalk going to become unsupported must be counteracted by the pertinent Organizations, like ESUG in Europe and STIC, and of course the Software companies themselves like VW and Instantiations. When I worked at Gartner, they had a similar table which included Smalltalk in the doom-way in "five years" ca. Y2K and recently they changed their mind a little bit: http://blogs.gartner.com/mark_driver/2008/10/09/remember-smalltalk/ and http://news.squeak.org/2008/10/11/smalltalk-is-cool-again-says-gartner/ (just a sample, Google is your friend if you're really interested).
OTOH, notice that another Research company whose business is selling 'insight' says something about the primadonna du jour: http://www.cincomsmalltalk.com/main/2010/12/java-is-a-dead-end-for-enterpris...
HTH
-- Cesar Rabak
Thing about these tales of "dying languages" & stuff like that is that it's really hard to avoid conflicting interests while judging things. I remember the golden age of Microsoft and its happy marriage with Intel, when companies (Gartner included) hammered execs minds with the TOC mantra (Microsoft as opposed to Unix/Linux - OSX not available at that time & Apple starving of Jobs). Following that line of reasoning, by now Oracle would be history, Linux would be history and Apple would be the only thing preventing Microsoft running into anti-trust laws. Obviously history unrolled different ways... I don't see smalltalk "retiring" or "going six palms down" because from the corporate vantage point smalltalk just didn't happen (yet). IMO, some things prevented smalltalk going main stream: (a) bad marketing from smalltalk players (b) Java happened and its shortcomings weren't evident to community from the start but Sun put a real bunch of money in it. (c) Microsoft had its own ideas about development platforms (C#, .net, etc) (d) Nobody presented a (really) suitable solution to large scale problems (performance, scalability, security, stability, etc), so corporations kept relying on "traditional" solutions (like C, C++, etc). (e) "Mainframe world" is a completely different breed. What's happening is that "software development world" is getting short of solutions to jump to next Fermi levels of systems development. Just a little look in industry and it becomes really visible. Some development cycles in "software consumer products" are exceeding five years & costs are surpassing hundreds of millions of US$. IMO smalltalk and/or its derivatives may fill important niches. But wether this happens or not will depend on several things. The most important of all is the availability of a clean open solution, coherent, supported and fully documented. Currently there are two divergent families of "open smalltalk": (a) gnu smalltalk (b) squeak/pharo/cuis Focusing in the squeak/pharo/cuis branch, I've noticed that pharo people got really concerned in aspects that can attract market interest. They're looking for funding, fighting to have "hired people" (meaning under wages) minding "hairy aspects" of development, maintenance, documentation and some sort of "standardization", etc. Collaborators also produced interesting "printed material" (books, tutorials) but IMO such material is still very academic in nature. People in industry likes best "manual" stuff (like "foundation classes" with "methods" (messages in the case) documented, etc). IMO (again) here lies a relevant problem: there are "no foundation classes" (no "landmarks" or things that cannot be easily changed) and it becomes apparent when people exchange ideas about "keeping this" and "getting rid of that". At least debate is going on and many issues have been addressed. One thing I don't (yet) like in pharo is that there's no easy way of applying fixes. Would be nice to have an update mechanism automating (at least) fixes installation. I think this kind of thing is important to market acceptance of smalltalk. CdAB