or How SAS Institute needs to take competition from WPS, (sas language compiler) in an alliance with IBM, and from R (open source predictive analytics with tremendous academic support) and financial pressure from Microsoft and SAP more seriously.
On the weekend, I ran into Jeff Bass, owner of BASS Institute. BASS Institute provided a SAS -like compiler in the 1980’s , was very light compared to the then clunky SAS ( which used multiple floppies), and sold many copies. It ran out of money when the shift happened to PCs and SAS Institute managed to reach that first.
Today the shift is happening to cloud computing and though SAS has invested 70 Million in it, it still continues to SUPPORT Microsoft by NOT supporting or even offering financial incentives for customers to use Ubuntu Linux server and Ubuntu Linux desktop. For academic students it charges 25$ per Windows license, and thus helping sell much more copies of Windows Vista. Why does it not give the Ubuntu Linux version free to students. Why does SAS Institute continue to give the online doc free to people who use it’s language, and undercut it. More importantly why does SAS charge LESS money for excellent software in the BI space. It is one of the best and cheapest BI software and the most expensive desktop software. Why Does the SAS Institute not support Hadoop , Map/Reduce database systems insted of focusing on Oracle, Teradata relationships and feelings ??
Anyways, back to Jeff Bass- This is part 1 of the interview.
Ajay- Jeff, tell us all about the BASS Institute?
Jeff-
the BASS system has been off the market for about 20 years and is an example of old, command line, DOS based software that has been far surpassed by modern products – including SAS for the PC platform. It was fun providing a “SAS like” language for people on PCs – running MS DOS – but I scrapped the product when PC SAS became a reasonably useable product and PC’s got enough memory and hard disk space.
BASS was a SAS “work alike”…it would run many (but certainly not all) SAS programs with few modifications. It required a DOS PC with 640K of RAM and a hard disk with 1MB of available space. We used to demo it on a Toshiba laptop with NO hard disk and only a floppy drive. It was a true compiler that parsed the data / proc step input code and generated 8086 assembly language that went through mild optimization, and then executed.
I no longer have the source code…it was saved to an ancient Irwin RS-232 tape drive onto tapes that no longer exist…it is fun how technology has moved on in 20 years! The BASS system was written in Microsoft Pascal and the code for the compiler was similar to the code that would be generated by the Unix YACC “compiler compiler” when fed the syntax of the SAS data step language. BASS included the “DATA Step” and the most basic PROCS, like MEANS, FREQ, REG, TTEST, PRINT, SORT and others. Parts of the system were written in 8086 assembler (I have to smile when I remember that). If I was to recreate it today, I would probably use YACC and have it produce R source code…but that is an idea I am never likely to spend any time on.
We sold quite a few copies of the software and BASS Institute, Incorporated was a going concern until PC SAS became debugged and reliable. Then there was no point in continuing it. But I think it would be fun for someone to write a modern open source version of a SAS compiler (the data step and basic procs were developed in the public domain at NC State University before Sall and Goodnight took the company private, so as long as no copyrighted code was used in any way, an open source compiler would probably be legal).
I still use SAS (my company has an enterprise license), but only very rarely. I use R more often and am a big fan of free software (sometimes called open source software, but I like the free software foundation’s distinction at fsf.org). I appreciated your recommendation of the book “R for SAS and SPSS Users” on your website. I bought it for my Kindle immediately upon reading about it on your website.I no longer work in the software world; I’m a reimbursement and health policy director for the biotech firm Amgen, where I have worked since 1990 or so… I also serve on the boards of a couple of non-profit organizations in the health care field.
the BASS system has been off the market for about 20 years and is an example of old, command line, DOS based software that has been far surpassed by modern products – including SAS for the PC platform. It was fun providing a “SAS like” language for people on PCs – running MS DOS – but I scrapped the product when PC SAS became a reasonably useable product and PC’s got enough memory and hard disk space.
BASS was a SAS “work alike”…it would run many (but certainly not all) SAS programs with few modifications. It required a DOS PC with 640K of RAM and a hard disk with 1MB of available space. We used to demo it on a Toshiba laptop with NO hard disk and only a floppy drive. It was a true compiler that parsed the data / proc step input code and generated 8086 assembly language that went through mild optimization, and then executed.
I no longer have the source code…it was saved to an ancient Irwin RS-232 tape drive onto tapes that no longer exist…it is fun how technology has moved on in 20 years! The BASS system was written in Microsoft Pascal and the code for the compiler was similar to the code that would be generated by the Unix YACC “compiler compiler” when fed the syntax of the SAS data step language. BASS included the “DATA Step” and the most basic PROCS, like MEANS, FREQ, REG, TTEST, PRINT, SORT and others. Parts of the system were written in 8086 assembler (I have to smile when I remember that). If I was to recreate it today, I would probably use YACC and have it produce R source code…but that is an idea I am never likely to spend any time on.
We sold quite a few copies of the software and BASS Institute, Incorporated was a going concern until PC SAS became debugged and reliable. Then there was no point in continuing it. But I think it would be fun for someone to write a modern open source version of a SAS compiler (the data step and basic procs were developed in the public domain at NC State University before Sall and Goodnight took the company private, so as long as no copyrighted code was used in any way, an open source compiler would probably be legal).
I still use SAS (my company has an enterprise license), but only very rarely. I use R more often and am a big fan of free software (sometimes called open source software, but I like the free software foundation’s distinction at fsf.org). I appreciated your recommendation of the book “R for SAS and SPSS Users” on your website. I bought it for my Kindle immediately upon reading about it on your website.
I’m a reimbursement and health policy director for the biotech firm Amgen, where I have worked since 1990 or so… I also serve on the boards of a couple of non-profit organizations in the health care field.
Ajay- Any comments on WPS?
Jeff- I’m glad WPS is out there. I think alternatives help keep the SAS folks aware that they have to care about competition, at least a little 😉
( Note from Ajay-
You can see more on WPS at http://www.teamwpc.co.uk/home
and on SAS at http://www.sas.com/
DECISION STATS 