Date: Tue, 30 May 2006 11:33:55 -0700
Reply-To: "William W. Viergever" <william@VIERGEVER.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "William W. Viergever" <william@VIERGEVER.NET>
Subject: Re: SAS as a programming language (OT: Response)
In-Reply-To: <200605300825.k4TAkHdC017569@mailgw.cc.uga.edu>
Content-Type: text/plain; charset="us-ascii"; format=flowed
gerhard - couldn't agree more
OK sherman, set the way back machine to .... late 70's? ... early 80's
mebbe bob hamer or nat wooding (or, perhaps, rick langston might
leave his nest for a minute) will pipe in here w/ a correction to me
foggy brain, but
i remember an ad (campaign?) by SAS that showed 3 different stacks of
IBM punch cards (that should definitely narrow down the time/date
range) needed to program, IIRC, a chart
also IIRC, i think the 3 stacks were for code written in SAS, PL-1 or
COBOL, and Fortran
yes grasshopper ... the SAS stack was by far the shortest
to this day, i think that that was one of the most compelling "ads"
i've ever seen
out
At 01:25 AM 5/30/2006, Gerhard Hellriegel wrote:
>That's a bunch of half-knowledge, what Lei Zhang wrote there. Nobody is
>forced to use SAS! Use JAVA, PERL, VB, C or FORTRAN (I assume you have never
>programmed anything more than "hello world" in one of that languages).
>What I'd be interested in: could you please show me the JCL-equivalent of
>
>proc print data=sashelp.class;
>run;
>
>a very simple SAS "program" if you agree...
>
>or maybe a proc reg, proc summary, proc gchart, ...
>
>Or show me a simple proc summary in FORTRAN!
>
>Or show me a user-interface, like you can have it with SAS/AF frames and SCL
>in a few minutes in one of the "good and modern" programming languages!
>
>
>
>
>On Mon, 29 May 2006 15:44:18 -0700, Howard Cherniack <cherns@COMPUSERVE.COM>
>wrote:
>
> >I can't see much JCL in SAS's PROC steps, except in a very generic
> >sense of "keyword=value". The Data Step language is definitely
> >modelled after PL/I, not FORTRAN--you want to see a FORTRAN-like stat
> >program, look at SPSS of the day, with specific columns for specific
> >items. Although object orientation wasn't yet dreamed of, PL/I was an
> >excellent and flexible procedural programming language, with free-form
> >syntax, and control structures (like DO and SELECT) that were far
> >ahead of its time.
> >
> >Knowing nothing about SCL (except that SAS/FSP costs extra, and none
> >of the places I've worked for have thought it worthwhile to buy it), I
> >have nothing to say to the main subject here, except to wonder what's
> >in Joe's responses that seem to be expressed in UUCP or MIME or PGP or
> >something. Cheers, --[the other] Howard
> >
> >> ------------------------------
> >
> >> Date: Sun, 28 May 2006 13:56:41 -0700
> >> From: Lei Zhang <lzhang9830@YAHOO.COM>
> >> Subject: Re: OT: Resposnse
> >
> >> I think I agree with Joe on that SAS uses an antique computer language.
> >> Its PROCs' syntax is borrowed from IBM Job control language (JCL) 50
> >> years ago, and its Data Step syntax is copied from Fortran language
> >> but without any user-define function capability up to now. However,
> >> SAS/SCL is not a langauge that is any better than SAS Data Step
> >> language. Thay both have been half-baked patatos for almost two
> >> dozens of years. There is no gain by switching to SAS/SCL.! That is why
> >> no one wants to pick it up. Before new serious competitors emerge with
> >> the support from US government agencies such as FDA, SAS Intitute will
> >> not make any signficant change on their langauges in the foreseeable
> >> future.
> >
> >> My two cents.
> >
> >> LZ
> >
> >Adp> ------------------------------
---------------------------------------------------
William W. Viergever
Viergever & Associates
Health Data Analysis / Systems Design & Development
Sacramento, CA 95825
william@viergever.net
(916) 483-8398
---------------------------------------------------
| 
