Putting unrealistic restrictions (only works in IE, can only use .Net) is going to hinder the progress of its development.
There are times where vetting and change management make sense. But then we have applications that were built before 2008 and realize that in the long run, with the tools of today, it would be easier to use the tools of today than adding another feature to this bloated mess.
We had services from ALL major IT companies in the world that still need IE (or Firefox, or Chrome, exclusively) , so this is even not strictly in gov or IE domain only.
Some gov companies are lucky to have somebody that lives in 21st century that is on such position that can change some of it but that is merely a fluke.
We had similar case - few years ago, the tax office switched to electronic VAT tax returns, which was Windows-only. There was a company that continued to fill their returns using the old-fasioned way (paper). In the court, they won, because the filing system was Windows-only, and the state cannot force you to purchase your software from a specific supplier.
Perhaps you can win on court. But what about time and energy you will lose. Nobody wants that. They got sued, thats what I am talking about.
And according to my Kibana presentation on IIS logs there are actually ~100 linux users in entire country (not counting mobile users on Linux of which there are a lot due to Android). This might be related to concrete gov domain (accounting and VAT) tho.
Its principles 100% for me.
[1] https://gds.blog.gov.uk/2012/10/11/what-devices-are-we-suppo...
I was at their Velocity conf presentation and immediately took them as show case on how gov stuff should be done.
We use Vue.js for frontends so 'using JavaScript very sparingly' doesn't even compute to me. Switched from Angular and I totally understand what you mean (given its more JS hardcore and more problems).
Any new projects should have a "migration" or "upgrade" metric associated - how long will it take to upgrade to the next major version of the employed platform? how long will it take to migrate to a different platform? no doubt there are not always simple answers, but it's something which needs to at least be considered. And there are sometimes really "single source" options with no viable competition, but those are possibly more rare than we'd think, and even if not, the risk needs to be called out even more in those project/contracts then.
This should be as plain and simple as stating that "if there exists a fully functional open-source technology that satisfies a public institution's need, the institution should be forced to use such technology instead of any proprietary alternative, or to demand it's contractors to develop a custom open-source solution, even if it's more expensive then a proprietary alternative" (with some "sane" limit like it can be "up to but not more than 5x as expensive" - there's befits to society from funding "a bit" of os dev from public funds, but not over do it...).
"Software capitalism" and public infrastructure don't mix well. And private companies can still make tons of money developing, customizing/extending and supporting open-source software for use by gov clients.
But if it's paid from people's taxes, it should only be open-source!
Yes, use oracle junk instead of postgre if you want, but software MUST BE open source, its only logical given that is financed by the budget and analyzed, tested and influenced by the gov people. Otherwise, you have private companies that get rich on taxpayer money, and not only that, but monopoly, zero reuse between regional local govs, worst quality and security etc.
You wouldn't believe the resistance I have, trying to introduce this in my country.
Half the stuff a government has contracted out for mils or bils could have been mostly implemented by small shops charging a few hundred K tops.
Unless you still believe a glorified CMS/ERP takes billions to write.
Given that, most would stay in gov I think - the domain is fantastic for serious engineers - entire country of clients, usually all domestic companies or people must be involved, most services should work 24/7, have zero downtime etc. Its as hard as it can get without being at Google and Facebook IMO.
There are other ideas to solve this problem too, but let me not digress here.
CITY: Hey, we paid for this, but these other cities using it aren't funding (chipping in) for further development...