4 ms·
Somewhat similarly, as a stopgap in '07 I built a Django (well really, Python and duct tape) Sarbanes-Oxley SOD system over a month or so. I had to deal with w
by gstar 17y ago
Somewhat similarly, as a stopgap in '07 I built a Django (well really, Python and duct tape) Sarbanes-Oxley SOD system over a month or so.
I had to deal with well over a million records, but fundamentally it was a pretty simple ask, laying a 2d matrix of rules over user and permission records from 30 different systems. I even hooked it up to SAP and various AS400s through MQ, print queues and various other things.
It was a total stopgap move - so a whole bunch of vendors bid on the "real" system, one was selected, and now almost 3 years later and a few million dollars spent, and the "real" system still isn't working.
"Web 2.0" frameworks have a lot of applications in business, the ORM + presentation layer tools give you most of the groundwork you need, but as the article says, the key is in the polish and the business oriented design.
- bendtheblock 17y agoI have recently begun a similar project: I work for a private bank and am rebuilding our client website. The main system (which is used internally for booking transactions and managing portfolios etc.) runs on an Oracle DBMS, so it isn't really legacy, but it does have a goulash data schema that is convoluted and slow. I'm using Rails for the site, which of course assumes certain conventions, but I've actually found it easy to loosely couple the site to the old system (by using views to extract the relevant data plus a little extra Ruby code). Most people in my dept haven't even heard of RoR but I've already had a positive reaction to the proof of concept which blows the old site out of the water and was built in less than a week. The UI advice in the article hits the nail on the head. When dealing with legacy systems, smart use of aggressive caching and AJAX (loading icons etc.) can help improve user experience without any changes to the underlying system.
- Retric 17y agoIf they have been using your system for 3 years, you might want to bid ~500k for a ~6 months contract to get it to the next level. It might not be fun but you can make crazy money building custom enterprise applications once you have some credibility.
- gstar 17y agoHa! Indeed! I would absolutely consider doing that if I had time. I'm kinda all-in on my own startup at the moment and I'm doing my level best to not be distracted by contracting. And oh how right you are about it not being especially fun! You make a good point about the sort of money that I should chase to do the job properly, too. Thanks.