4 ms·
Gotta be the most archaic language still in common use. I mean not saying it doesn't have legitimate uses, fast well tested and a few procs with the right opt
by analogic 9y ago
Gotta be the most archaic language still in common use.
I mean not saying it doesn't have legitimate uses, fast well tested and a few procs with the right options can accomplish a the equivalent of a lot of Python, and I only ever learned enough to run a few datasets and port some old code to python/r/vba, but observing the crazy ass shit people did to implement any kind of logic or data transformation was enlightening. Many many "wait people actually do this" moments. Like an entire ancient programming paradigm I never knew existed.
- forkandwait 9y agoSAS actually combines several useful things in one seamless environment : statistics, plotting, ETL, and data management. For any one of those functions, there is some better system (R, R, perl, sql respectively), but integration can be hard. Also, it is a single purchasing decision with a standard product. Plus, a good SAS programmer can build fairly big processes and avoid brain dead IT bureaucracies. Not that I love it, but it's a niche no other single environment serves.
- dredmorbius 9y agoIt's not so much that integration is hard, but that any given site tends to integrate differently. The problem I've encountered with SAS specifically is that the more powerful tools are hard to gain access to (or proficiency with) given the licensing costs. I have literally been told by SAS regional sales staff that I would be better off pirating a copy of the software to gain access to it. I politely declined.
- dredmorbius 9y agoAntecedents from, as noted, mainframes, and APL specifically.