#SAS C Upgrade $$$
30 messages in this thread
Hi,
Just curious and wanted to take a poll about how SAS C users feel about
the upgrade price of $109.00 (plus local sales tax) to go from version 5.xx
to version 6.x ? Admittedly, to buy in new will cost $395 but…whew…$109
to upgrade…I dunno. Maybe DICE is looking better all the time. Any
thoughts/opinions/etc. appreciated.
JeffS
I think the price is in line with other C compiler upgrades… It just cost
me $120 to upgrade my Microsoft C package and well over $200 to upgrade my
Borland C++ package…
For 1200 pages of printed documentation (a complete rewrite over the old
manuals) and the work that's gone into the new tools (and the online help)
– yes I think it's worth it!
…Steve
Ok,
I just wanted to make sure that something worthwhile was being added
beyond a pretty GUI (that I'll never see/use being a firm CLI user). The
advertisement didn't really go into depth and as a long time Lattice/ SAS C
user (with plenty paid into past upgrades), I just wanted to be sure the
$109 + tax would be worth it. Perhaps some insights could be given as to
improved code generation, etc. The 'Facts' sheet provided with the upgrade
notice only seems to tout how much more 'user friendly' the system is and
not the performance aspects.
Jeff
The two performance improvements that impressed me were the Global Symbol
Tables, and the lack of LINK and UNLK instructions for every function.
With a GST that has <proto/all.h> in it, you can #include <proto/all.h> and
not even have to wait the 5 seconds or so it takes to load it in as
Precompiled Headers – it stays resident in memory until flushed (either
manually or by a memory panic).
Actually another nifty thing is that the compiler itself is not a
collection of executables anymore – now it's "lc1.library", "sc2.library"
(oops change that to sc1.library) with the SC program calling the libraries
to do the compile. So if you have the memory, the compiler is
automatically made resident and if you run out of memory, the compiler
automatically vanishes.
…Steve
Steve, you have to remember that in order to get all of that performance,
you have to have GOBS of memory. On a smaller (2 Meg machine) you can't
effectively use the GSTs and you lose quite a bit of the performance gains.
Even with the library approach, the compiler has a tough time fitting on a
floppy system.
True. But anybody spending $300 on a compiler probably already has a hard
disk – anybody seriously using the compiler will have one. I would rather
see the compiler optimized for the high end.
Floppy users can still use compacted headers copied to the ram disk, and
the compiler will load off disk just like it always has. But folks with
the extra memory and hard disk space get a much better compiler.
…Steve
>2MB of memory = GOBS? 🙂 I'd say that's darn near bare minimum for
productive Amiga use! 🙂
Dan McCoy – via Whap! and saving money doing it too!
That's not true, John. Global Symbol Tables that are created
specifically for one project are a few hundred K in size.
A two-meg machine would certainly have room to keep a GST
of that size in memory.
If you create a GST containing ALL the header files, it is
pretty big (~700K), but no project that I know of would
use all of them.
–Doug
Hi Steve,
Sorry I didn't reply earlier but I live in Miami and we had this
small problem with the weather earlier this week 8-). I just got power back
3 hours ago. I suppose I will upgrade to 6.0 eventually but as Steve
Ahlstrom pointed out, I was initially in a state of shock also at paying
almost 50% of the original price for the upgrade. I guess my wallet talked
before I thought things out. Thanks for the information concerning 'hard'
improvements.
An Andrew Survivor,
JeffS
Jeff,
I paid $205 at an Amiga Store in Feb. of this year for 5.10. Six months
later, SAS wants $109. That is 50% of my original cost and leaves a sour
taste in my mouth.
Allan Baer ( Be part of the solution, not part of the problem. )
And if they'd been blowing it out at $50, then the upgrade owuld be double
the price you paid for the original.
Fact is, SAS put more effort into the 5.10->6.0 development than most Amiga
companies put into an entire product. Of course they have to charge for
it.
…Steve
Do you work for SAS? Where do you get your "FACTs" that "SAS put more
effort into this development than most Amiga companies put into an entire
product" ??
Allan Baer ( Be part of the solution, not part of the problem. )
I've been beta testing 6.0 for quite some time now. The code generator in
6.0 is basically a rewrite from 5.10 – and it shows. SAS doesn't just do
the compiler for the Amiga – so the development they put into the compiler
on any platform shows up on all the others. SAS/C 6.0 is the result of a
lot of people working on the compiler for a long time.
(It's not a verified "Fact" because nobody has done any research into this
kind of thing).
…Steve
Steve,
I have spent the weekend messing around with DevKit. This program
works with CED and the AutoDocs. The program was done for WB1.3, but the
new 2.0 AutoDocs do not follow the old conventions. I've gotten the doc
files to convert and work and was starting on the headers; this looks to be
a little above me.
I now notice that SAS6.0 uses some type of on line help. Will this
do structure look up? Will it aid me in finding what headers to use in a
program? Or is this help just "C" function help?
-Tom
The SAS/C help is for stuff SAS provides only. Standard IO functions,
character functions, that sort of thing, as well as helpw ith compiler
options, the utilities (Make, etc), the debugger, the editor, and probably
some other things I've forgotten.
…Steve
Thanks Steve for the insight on SAS/C help. I have gotten the DevKit to
work satisfactory with CED. So I'm in better shape then when I asked the
question.
-Tom
Ah, but if you are using the Global Symbol Tables, you can browse
structure definitions, preprocessor macros, typedefs, function
prototypes, you name it. There's a special utility provided
to do this.
–Doug
Actually, it has not been true for over a year that the SAS Compiler work
was coming from another platform. They switched all of the internal
machines from being 680×0 based HP 425s to be the new HP PA-2 RISC
platforms. The compiler front end is also not the same as any of the other
SAS compilers and the Backend is Amiga specific (or at least it was when I
left there a year ago). There were only 3 people working on the entire
Amiga product at that time. (I have been avoiding these discussions except
when someone posts something that is inaccurate).
NOTE: I have NO association with SAS Institute.
>There were only 3 people working on the entire Amiga product at the time.
Which I should add is not necessarily a low number.
But when the work stopped on the compiler, the 5.10 version was still the
current Amiga release, right? How many people were working on the code
generator and optimizer and debugger, which did come from other SAS
efforts?
…Steve
Sorry, John, you are incorrect again.
There are considerably more than three people working on the
Amiga product. We had four developers working on the early
stages as long as three years ago… the fourth one being
someone I'm sure you can recall if you think hard enough.
We have had three developers on basically JUST the Amiga
6.0 product for most of the past year. During that year
we also had two tech support people devoting a good portion
of their time to 6.0; we had three testers spend a total
of about 1.5 man-years on 6.0; we had two technical writers
several editors, proofreaders and so forth; we had several
people in marketing and sales areas spending some time on
it.
My personal estimate is that there is about 15 man-years worth
of effort in SAS/C 6.0, 7 of which is specific to the Amiga
product and the rest is portable.
–Doug
I work for SAS Institute.
I'll be more than happy to state that we've put more effort
into this than most Amiga companies put into an entire product.
SAS/C 6.0 is virtually a ground-up rewrite of all the major
portions of the product. And it's a HUGE product compared to
just about anything else on the machine, with the possible
exception of AmigaDOS 2.0 itself.
Steve does know what he's talking about – he's lived with the
product through the beta and gamma test cycles since November
last year.
–Doug
Version 5 is four years old. In that time SAS Institute has asked
for money only once, for the 5.10 upgrade, and that was only $40.
Somebody has always just purchased the product when an upgrade
goes in, there's just no way around it except not to upgrade.
You still have the option of not upgrading and sticking with
5.10. 5.10 is as good now as it was before we announced the
upgrade. I hope that you'll agree that the upgrade is worth
it once you see it.
–Doug
It all depends upon how the price translates, for people outside the US.
All too often, a RKM selling for approx $35 in the US ends up in England
for 35UKP, not far off a doubling in price. I would be happy to pay the
equivalent of $109 plus any local markup, but >100UKP would be getting
hairy !
Steve the G (Steve Goddard 100014,674) [BEDFORDSHIRE, UK]
Considering you get new manuals and a quick reference manual along with the
multitude of disks, $109 doesn't sound unreasonable to me (even while being
unemployed). I've had more costly product upgrades like Imagine 1.1 to 2.0
that cost more and I got less out of it.
Dan McCoy – via Whap! and saving money doing it too!
It does seem a bit pricey, but considering the new price when I paid for
lattice C 5.00 at 495.00, I do have one problem and that is I can't seem to
get SAS to realize that I purchased C when Lattice distributed it I have
all my registration numbers and I have sent many letters to SAS all have
been unanswered. Does SAS have a Forum here on Compuserve.
Joe,
If you paid $495 for SAS/C 5.0 you got taken. I believe that suggested
retail is $295 (I paid $217 plus tax for it).
-sja
No, SAS Institute doesn't have a forum here.
Please call our sales people at (919)677-8000. I'm surprised
to hear that you have been having trouble getting letters
answered, they are usually very good about keeping track
of correspondance. Just to make sure you have the correct
address:
SAS Institute, Inc.
SAS Campus Drive
Cary, NC 27513
–Doug