serverside COBOL is still going strong. So what was your point?

I thought it was clear - it has fuck all to do with the merits of the various bits and types of java - its lost the battle for popularity. Its seen as insecure and slow.

Nobody wants the runtime on their systems, so any java app will not be as popular as any written in a different system.

for this week....

Could be that was the start of the decline, but I think the real killer blow was when all the browsers decided to put it on the equivalent of a 'do not fly' list.

Everyone and his dog now knows its not to be trusted, its a security nightmare, in addition to being dog slow and having really poor UIs.

(whether that is true or not, doesn't matter).

So now.. who wants to be a Java developer? Its akin to admitting to work for Walmart in the personal hygiene section.

Nothing stopping the open source community from implementing a 100% compatible .NET library. Other than not giving a shit about it.

you want .NET on Linux, go ahead and port it yourself. If its so easy then you should have it done by teatime. Nobody in the FOSS community wants it to be ported, and they don;t want to use it either, so there is no expectation of such things happening.

you forget... it'll never be released. Its job is to garner support, "hearts and minds" and then get all the best bits subsumed into the core of Microsoft closed-source products where you'll never see it again.

then Roslyn will not be needed, can be left to die while they produce another open source project.

Indeed. Apart from the maths, the software side of the course used different languages for different aspects.

For example, I was taught threading using Concurrent Euclid. We use Prolog, Simula, assembler and other esoteric languages for 1 term at a time.

The cynic in me thinks a lot of the move to Java as a single language to teach was partly driven by the idea that graduates should have learned things immediately useful to industry, who was touting Java at the time.

heh. When I did my computer science degree we were taught more about statistics than software.

At the time I figured what a waste of time, and for a few years it was right... until we got to the large scale systems where job throughput matters quite a bit. Then all that old crap turned out to be useful. Who'd have thought modern, super-fast computers still have the same limitations as the ancient mainframes! (well, with a lot more users, admittedly)

My suggestion for that rounded education is to learn some assembler. C is a wrapper around a lot, and to be fair is the minimum you can get away with for writing programs in the real world. But to truly understand what that computer is doing, you need to be exposed to the accumulator and program counter, then you'll get the epiphany that computers are incredibly, totally stupid.

You'll also get an insight into the mass of stuff the higher level languages do for you, and when enhanced with a framework (hey, or two!) and some libraries and so on, you'll realise what tortures you're inflicting on a CPU who's main job is to add the contents of 2 memory cells together.

depends. Did they get a bonus for getting the takeover through?

Takeovers tend to be very popular for just that reason.

you missed out the bit where HP wanted the Automony enterprise search stuff so badly, they just didn't bother working out how much the rest of the business was actually worth and made them a stupid (by Autonomy standards) but ordinary (by Silicon Valley takeover standards) offer.

Mind, with WhatsApp being bought for $16bn, and only bringing a free> product to the table, you'd think the Autonomy deal was the paragon of financial prudence. that HP turned it into the equivalent of a free app is down to the competence of HP's management.

Oh yeah, you already covered that. Was worth repeating though.

Except that in the UK we do have freedom of speech - just go to Speaker's Corner in London and hear all the nonsense spouted by idiots with all kinds of conspiracy theories.

None of them are doing it anonymously.

Many others in the UK have spoken out with dissent for the government's intentions. Sometimes it works and they change, most of the time they just ignore us and do what they were going to anyway.

I think the difference is that, even though you're not truly anonymous online either, too many fuckwits think they are and it affects their behaviour. Look at the recent twitter arrests and imprisonments for doucebags who thought it was ok to spout death threats online.

So all they're really saying is that you should be more open about who you are online, as the authorities can find you anyway if they dislike what you've said. Removing the appearance of anonymity may make those morons change their behaviour online, and will make finding you cheaper when (or if) you say things worthy of police action (I assume if you're worthy of special branch attention, they will already know who you are with a quick call to GCHQ)

Yeah, so I hope and expect Jesse Jackson to really kick off and promote the workforce equality such that it is truly representative of the local population's ethnic ratio.

If that means more white workers have to be hired then I expect him to be promoting more white guys.... won't hold my breath waiting for that to happen though!

you'd have a vast library of libraries. Something like CPAN or something you'd get in the C world. Libraries written to perform some task and nothing more. Then documented with care and the API published.

Anyone wants to do something, they take the library that appeals to them and adds it to their program and build up a program from these bits.

Now the problem today is that a) some only use libs that come with the OS or language framework, b) the libraries that are out there are shit, written quickly and for a bit of a mishmash of scenarios.

For example, you can get an XML parser and it will work perfectly. It will only parse XML though, but then, that's what only what you want from a XML parser library! .... well, except it also comes with network routines and specialised SOAP parsing and a suite of http helper methods, including a web services subsystem :-(

So the problem is not so much that we have libraries, but that the libraries we have are not good enough as library code.

I'll agree there - thought its not Java at fault necessarily - not unless you lump in a bunch of other languages like VB, C#, JS etc.

The problem is of the library code you're using. Libraries should be small, well defined, easy to use, and documented.

The problem today is (especially with code written in Java, .NET or JS) that it is knocked up to solve some problem but the problem is not only not properly understood, but the code that is provided doesn't solve it particularly well. Its not defined as a discrete task, more as part of some greater whole that someone thought "worked ok for me in my circumstances, so should be fine for others too".

If libs were properly specified as libraries and their API documented fully, then we would see more code reuse and better, cheaper code. If only, but the cost of making such a library tends to be too slow and difficult for the 'I want it now' majority, and this is why we continue to have this kind of shitty code problem.