leisner.henr801c@XEROX.COM (06/14/91)
Jim Showalter said: You get what you pay for. My professional experience in software extends backwards a decade. I've used CP/M, MS/DOS and Unix machines. Basically, software with source code is generally of higher quality to use than random software. The moderated software I've seen on USENET and the gnu software is of extrodinarily high quality. One reasons many vendors don't provide source code with programmers tools is they're probably ashamed of it. I find the GNU stuff to be of much higher quality then anything else I've seen (I use gcc extensively) and its free and comes with source code. Which gives me a logical process if it does something wrong, where I can elect to fix it or report the problem. I've seen a decrease of the quality of software in the industry over the last 5 years (on CP/M and early Ms/Dos everything I tried worked pretty good). Now a lot of products on MacIntoshes and PCs do "magic" without documenting what they're doing or providing feedback to the user. While the quality free software is being distributed, used, improved upon. The DoD probably made a mistake for providing a framework for a vendor system rather than taking responsibility for a standard ADA compiler itself amd distributing it cheaply. What I've found is software vendors are in business to sell a product and will only fix problems if enough of their customer base complain. Free software writers fix problems all the time and release new versions much more quickly than vendors. So in this strange business, "you get what you pay for" doesn't hold true with software tools. marty (Knowledge is useful in the Information Age) (Software is mindstuff. It is the hardest activity created by man) ARPA: leisner.henr801c@xerox.com NS: leisner:henr801c:xerox UUCP: hplabs!arisia!leisner
rgc@raybed2.msd.ray.com (RICK CARLE) (06/20/91)
In article <"13-Jun-91.14:53:27.EDT".*.Martin_A._Leisner.Henr801C@Xerox.com>, leisner.henr801c@XEROX.COM writes: > > The DoD probably made a mistake for providing a framework for a vendor system > rather than taking responsibility for a standard ADA compiler itself amd > distributing it cheaply. > I agree that expensive software is often crap and free software and shareware is often great stuff. But I certainly take issue with the above statement. The Army's decision to drop support of ALS (in 1986 or so) and encourage an open Ada market was no mistake - it was brilliant. It was a key element leading to the success of Ada (of course Ada is successful). ALS was to have been the Army's standard Ada compiler. But in 1985, it was only compiling about 20 statements per minute. Development was lagging far behind schedule. Then DEC released VAX-Ada, which ran ten times as fast, was the most robust Ada compiler seen to that date, and produced code that was comparable (in most cases) to its FORTRAN compiler. Suddenly, the light dawned: the computer and compiler industries could easily beat government-sponsored compilers just by responding to market demand and leveraging (reusing) their existing product base. The brilliant part was not in recognizing the benefits of a competitive market (that had become apparent to everyone), but in having the courage to change policy - a rare virtue for governments and their agencies. In August 1985, there ware 22 validated Ada compilers; in November 1990, there were 318. This growth is undoubtedly due to the open, competitive Ada market. It was the right decision for its time. Perhaps now the situation has changed a little. Compilers are plentiful, but they're too expensive for many universities. Maybe now is the right time to revisit the concept of a DoD-sponsored cheap compiler. Sponsor a competitive procurement for inexpensive Ada compilers to host on easily available college computers: PCs & Macs. Buy one super-site license and distribute them to any college with an Ada course. Just wishful thinking. Rick Carle
emery@Dr_No.mitre.org (David Emery) (06/21/91)
Probably the greatest benefit from the ALS and AIE government-funded Ada compiler contracts is the large number of people who worked at SofTech and Intermetrics, and then went on to build a second (successful) Ada compiler (usually somewhere else). This is a side of Ada technology transfer that has not been adequately recognized and acknowledged as part of an Ada success story. dave emery
brett@tle.enet.dec.com (Bevin) (06/21/91)
In article <EMERY.91Jun20145620@Dr_No.mitre.org>, emery@Dr_No.mitre.org (David Emery) writes... >Probably the greatest benefit from the ALS and AIE government-funded >Ada compiler contracts is the large number of people who worked at >SofTech and Intermetrics, and then went on to build a second >(successful) Ada compiler (usually somewhere else). This is a side of >Ada technology transfer that has not been adequately recognized and >acknowledged as part of an Ada success story. > > dave emery I don't know about the other compiler vendors, but I suspect that the major ones [DEC, Rational, Verdix, Alsys...] and many of the others [...] had already completed their first Ada products by the time the large numbers of people from SofTech and Intermetrics came looking for jobs. I know that was the case at DEC, and are fairly sure about the other three I named. /Bevin
pierson@encore.com (Dan L. Pierson) (06/21/91)
In article <2356@raybed2.msd.ray.com> rgc@raybed2.msd.ray.com (RICK CARLE) writes:
It was the right decision for its time. Perhaps now the situation has
changed a little. Compilers are plentiful, but they're too expensive for
many universities. Maybe now is the right time to revisit the concept of
a DoD-sponsored cheap compiler. Sponsor a competitive procurement for
inexpensive Ada compilers to host on easily available college computers:
PCs & Macs. ...
There may be a cheaper and more productive alternative to yet another
heavily government financed boondoggle. Current DoD validation
requirements have the effect of prohibiting free or even cheap
validated compilers because the validation costs can only be born by
relatively large organizations that charge large prices*.
Could a combination of subsidized validation for a few free or low
cost compilers and maybe a small amount of carefully selected grants
help produce the cheap Ada compilers that universities and private
students need?
*I get the feeling from comments about the value of "real commercial
compilers" that some people in this group believe that this is a
feature :-(. Maybe this view is logical if you believe that only
"real software developers" should use Ada and that "real software
development" can only take place in large, well funded organizations.
I think that such attitudes run directly contrary to the history of
economic growth and innovation in the USA; most real growth and
innovation come from small groups, mostly small businesses. Making
Ada effectively unavailable to such groups is helpful neither to the
groups nor to our software industry as a whole.
Aside: It's rather odd to find myself, an Ada opponent, supporting the
language this way. But its basic design is much better than the main
alternative, C, and we badly need to move most of our work from C to
real high level languages just as we finally moved from assembly to C
quite a while ago. It's a real (expensive!) pity that none of the
language with most of Ada's advantages but without its bulk and
misfeatures seem to have a chance.
--
dan
In real life: Dan Pierson, Encore Computer Corporation, Research
UUCP: {talcott,linus,necis,decvax}!encore!pierson
Internet: pierson@encore.comjls@netcom.COM (Jim Showalter) (06/22/91)
>Probably the greatest benefit from the ALS and AIE government-funded >Ada compiler contracts is the large number of people who worked at >SofTech and Intermetrics, and then went on to build a second >(successful) Ada compiler (usually somewhere else). This is a side of >Ada technology transfer that has not been adequately recognized and >acknowledged as part of an Ada success story. Yet more proof that one always builds one to throw away. Nothing wrong with this, provided it IS thrown away (which brings up ALS/N...), and provided it doesn't cost too much, which I think ALS did. Hell, Rational built its entire environment for less than the government spent on ALS. -- *** LIMITLESS SOFTWARE, Inc: Jim Showalter, jls@netcom.com, (408) 243-0630 **** *Proven solutions to software problems. Consulting and training on all aspects* *of software development. Management/process/methodology. Architecture/design/* *reuse. Quality/productivity. Risk reduction. EFFECTIVE OO usage. Ada/C++. *