ElephantBanjoGnome wrote:
Certainly in the public sector, I.T.-ignorant management types draw a direct correlation between cost and perceived quality. The most recent IT bungle in the organisation for which I work involved paying in the region of £2m for payroll and purchasing software that didn't work, wasn't well integrated, wasn't maintained, and eventually junked after 18 months of hassle. Authorised, of course, by a higher-up with no fucking clue about anything. For a tenth of that you could have paid a team of 5 well-paid developers for a year to come up with something better and more bespoke. They never considered that, naturally.
Oh God, so very much
.
I work in the dedicated IT department for a government agency and over the years we've built many bespoke systems which were only as complex as they needed to be to meet the business requirements of the time (while still being scalable and extensible). All systems were delivered relatively quickly, relatively cheaply, to a high degree of quality and all integrate well with each other. Most importantly though we can support them well because they're only as complex as they need to be and we know how they work; having built them from the ground up.
A couple of years back we needed replacements for two very old legacy systems, plus the business wanted to beef up those functional areas and better integrate them with existing systems. Shall we go with our proven track records of building supportable bespoke systems that fit the business need? Oh God no, let's toss that idea in the fucking bin and buy in some off the shelf packages that we'll customise, with the 'assistance' of third party consultants naturally. Predictably, both projects were a fucking disaster. Both base systems were stupidly complex over-engineered pieces of shit that barely did what they were sold as doing; much less what we wanted them to do. Unbelievable amounts of money were spent and the result was that one system went live and literally fell over the same week due to terrible performance, while the second system only made it into production recently (about two years late) and is running in parallel with the legacy system it was intended to replace because we can't fully trust the replacement system yet. The one that fell over was yanked from live and we ended up kicking out the third party dev team so as to redo it ourself with the assistance of the actual vendor. To get the thing running at an acceptable pace we ripped out gobs of functionality and it's now a shadow of the design we were originally promised would work. Oh, and it constantly suffers from mystery performance issues where it'll slow to a crawl for random amount of time and the vendor themselves are blatantly running out of ideas as to what the cause is. The upshot of all this is many, many horrible headlines and our hitherto very good reputation flushed down the fucking toilet due to shite management decisions that we as a department had no input into.
So, a possibly worthwhile experiment from which surely we'll have learned some lessons eh? Surely, now that another of our core systems has reached end of life (it's running Oracle 8i!) we'll go back to our previous and very successful delivery model? No. Of course we fucking won't. Our senior management and the goverment department that have oversight on us want to buy in an off the shelf product and customise it. Gnnnnhhhh! *bangs head on desk*