Why? On the intranet you're likely dictating the browser/technology used anyways. Why not simply use what's best for the specific job. If it's Flash, Silverlight, HTML5, rich client app, virtual machine, etc...
So I should not use the best tool for the job, and give up say 5x productivity, in the fear that some company will take me hostage -- especially since AFAIK, this has never happened. And if it has happened it has happened an order of magnitude less frequently as projects getting killed by using the wrong tools.
That I've seen happen first hand, including as recently as last year.
I've been writing a Silverlight app at my job for the past 8 months. I don't feel like Silverlight has given me any kind of productivity boost. I actually find Silverlight to be rather heavyweight and slow going, especially RIA Services which is boilerplate central.
I may have a bit of a "grass is greener" outlook, but I do feel my productivity would be higher with a web framework such as RoR. Especially if I was just as good at Ruby/RoR/JS as I am now at C#/Silverlight (I've been using .NET professionaly since its debut)
Then don't use Silverlight. My point wasn't to use SL per se. It was to use the best tool for the job. If you're most productive writing in 6502 assembly and then writing a binary rewriter to convert it to Logo, which you then compile to Javascript, go for it.
Did you consider all factors, when you are so sure about it being the best tool for a job?
Few years ago, there were many people absolutely sure, that creating IE6-only webapps is also best decision. And look today, where that decision got them.
Then don't write any code, because clearly you don't have enough belief that you (or I or whoever) can weigh the various tradeoffs to make the right decision.
Too often the decision masquerading as "best tool for the job" is in reality "the technology I'm fan of".
I really would not want to be the person, who few years since now is responsible for maintenance of some legacy Silverlight application. It is going to be exactly ugly as maintaing IE6 only application today. The difference is, that it is easier to avoid that now.
I really don't know what would make Silverlight best tool today (except for very specialized solutions). No, support in VS does not count.
Well then what would you use? HTML? That's what the guy who said to based their whole new system on Gopher said in 1992.
It's a stupid argument. I mean really a STUPID argument to say "don't use the best tools, use the tools that some random guy on HN says to use". If you can't trust your own judgment I'd quit software development right now. You really just have no business writing code. Because if you can't pick the best tools based on info, how can you pick the right architecture, or pattern, or prioritize which bugs to fix, or which scenarios to support, or which person to marry or when to have kids or when to buy a house.
I hate to be so blunt, but this reminds me of religious zealots who say, "ignore your best evidence... just trust us". Anyone who goes down that path simply has lost all of my respect.
And yes, downvote away. I just have trouble believe someone on a hacking site would say, "don't do what you think is best based on your own due diligence... follow some arbitrary crowd".
As long as your users don't want to use most mobile devices. I'd much rather put out HTML5 apps where possible... intranet apps in particular seem to tend to live for longer than you ever thought possible. But, yeah, it's definitely nice for certain scenarios.
