Janus ADA

The following are the responses regarding my request for information on the
JANUS/ADA compiler.


From roberts@virgil.UUCP


The $99.00 JANUS/Ada compiler represents the low end of the RR Software's
line of Ada compilers. All have been validated. Here's a rundown on the
line:


JET Ada Tutorial $99.00 produces .COM files only limited library
Educational Pak $395.00 produces .COM or .EXE files... larger programs
Development Pak $799.00 includes assembler/disassembler and full library
Systems Pak $7500.00 source code to run time system.


I've been using the development pak on my 8-MH/ 65-microsecond hard disk
system at home and a 12-MH Compaq 286 equipped with a 25-microsecond hard
disk at work. Waiting for the slower system to compile and link even a
small program is frustrating. You will need over 575k of memory. (Use of
an 8087 math coprocessor is optional.)


With the $99.00 compiler you will be able to evaluate RR Software's
compiler and documentation. You should be able to do some basic things
but not as much as you could do with one of the more expensive compilers.
If you elect to upgrade, you get credit for purchasing one of the "lesser"
compilers. I've been an RR Software customer since they put out a
subset compiler for the Osborne-1 CP/M base machines. RR puts out a
good (not a great) compiler.


-----


From bgun@trwind.ind.TRW.COM


My employer sprang for the $99 JANUS/ADA based on the the ad in BYTE magazine.
First: It appears that it really is a validated ADA compiler.
Second: The ad is extremely misleading and only one step above "slimeball".
It is relatively easy to figure out who the other 2 compilers they are
comparing themselves to from the information provided in the ad.
But what you get for $99 is not in the same league as what they compared
themselves to. On page Inst-3 it states:
      "Only memory model 0 (the smallest model) is available in the C-Pak. The
      bigger packages support more than one memory model. ... Programs are
      limited to 64K of code; if more is needed, we recommend an upgrade to
      an ED_Pak."


So for $99 you get a small model compiler and this in itself would be a
very difficult shortcoming to get around if you were planning serious program
development when you bought the package but based on the verbosity of ADA
I think it would have to be one hell of a compiler to be useful for anything
other than the most trivial of tasks.


Of course they do offer you the chance to upgrade. But then you are paying
the same price as the compilers they compared themselves to in the first place.


I have not had a chance to find out just how good (and compact) the code
generated by this compiler is and to be quite honest, I probably never will.
We'll just write this one off as a bad investment and go get one of the real
compilers.


Maybe if they had been more honest up front I might have considered buying
their compiler along with my consideration of the other ADA compilers on the
market, but you know what they say: "Fool me once shame on you, fool me twice
shame on me." I don't think I will give them that second chance.


-------


From: kent@xanth.cs.odu.edu (Kent Paul Dolon)


I've seen this in Janus advertisements (Computer Language Volume 5 Number 5,
Page 54 full page ad: "... Validation Suite ACVC 1.9") and I doubt the AJPO or
whoever is in charge of the watchdog process would let this go on long if it
weren't true, but I don't want to be the one to furnish the authoritative
"yes"; I don't have the right.


Remember that "validated" means it passed the validation test suite; it does
not mean "useable", "fast", "friendly", "produces efficient code", or even
"bug free", so check it out with some current users before taking the plunge.
(The benchmark performance stats in the above mentioned ad are _very_
impressive against the competition, though; hard to tell what the trade-offs
are.)


All the mandatory "buyer beware" items taken care of, however, you have to
admire the persistence of the Janus folks. Back when Ada/Ed was the only
validated anything, and the whole universe was saying a working Ada compiler
was never going to happen in our lifetimes and was beyond the state of the
compiler writers' art, and subset compilers were "illegal", and delivery dates
on Ada compiler contracts were slipping a year and more, an awful lot of
government contractors' start-up and proof of concept Ada programming
contracts were being done on humble Janus/Ada
subset-of-something-or-other-you-guess compilers running on Apple-II+'s with
CP/M cards. The Janus folks have stuck with the project for at least four lean
years, and I think at least one corporate transition; if they have finally
been validated, it is hard to find a group more deserving. If they want to
continue producing the poor-man's Ada, more power to them. Whatever its
performance, I'll bet it is worth more than 1/30th of the $3000 product at the
high priced end of the personal computer Ada compiler market, and if there
were an Amiga version at that $99 price I'd snap up a validated one in an
instant, for me.
--
John Herman ARPA: herman@marlin.nosc.mil
--

Return to the comp.compilers page.
Search the comp.compilers archives again.