Re: Ruby and .NET

On Sun, 2002-02-17 at 18:45, Sean Russell wrote:
> Sean Middleditch wrote:
> 
> > Thus, hopefully, an Open implementation will be among the best.  ^,^
> 
> Absolutely.  I guess we only differ in opinion about what the value of that 
> "best" is. :-)

Aye, suppose so.  ^,^

> 
> > Well, I was thinking more like how ORBit would still be useful to GNOME
> > even if it couldn't talk to another vendor's ORB.
> 
> Yes, again, good point.  My answer to that, then, is that most ORBs add a 
> certain overhead which often isn't worth-while unless you plan on 
> connecting with the outside world.  While I'm not positive I agree with 
> them, the KDE team had a point in choosing an IPC mechanism closer to the 
> metal (and avoiding the at-the-time heavily buzzword laden CORBA), in that 
> CORBA added overhead and complexity that degraded the end result, rather 
> than enhancing it.  I have to say that I don't see anybody really taking 
> advantage of the fact that Gnome ships with a CORBA ORB.

Which is unfortunate, actually... ~,^  Not taking advantage of
technology that is there is a real mistake, in my opinion.

> 
> >> change".  It isn't in their best interest, financially, to help Linux,
> >> and MS is not known for their altruism.
> > 
> > Well, not only MS will be selling these services.  The idea behind .NET
> > is that *anyone* can start a service.  WIth a functional implementation
> > of .NET on Linux, Solaris, BSD, Mac, etc., you can bet a good number of
> > services aren't going to be running on Windows servers.
> ....
> > I don't see Microsoft forcing these services to shut down.
> 
> Perhaps.  Again, I just find it hard to believe that MS is encouraging 
> anything that might increase /competition/. I fully expect that they'll 
> encourage /innovation/.  .NET services which are popular, they'll roll out 
> their own versions of which will, miraculously, work better and faster, and 
> soon after that, the original service will begin suffering mysterious 
> incompatibilities with the infrastructure.  The original company will have 
> to use an increasing amount of resources to resolve these problems, being 
> able to put less money into new development -- I'm not being terribly 
> anti-MS here.  This is typical of their behavior in the past, and it is why 
> they've been so successful.  Don't think it won't happen on .NET.  History 
> shows that the playing field is not remotely even when MS controls the 
> platform, and .NET is a platform.

The funny thing is, in most cases where MS tried to win a market,
they've lost.  Quicken is still a lot more popular than MS Money, AOL
more popular than MSN, and so on.  In the cases where Microsoft *did*
win (web browsers, media) Microsoft simply had the better product.  IE
is a hell of a lot better than NS4.x (6.x just recently jumped on the
scene, so far as the average user goes), and I'd use WMP over RealPlayer
any day.  I generally don't stick the word innovation with MS, but they
have mastered the art of improvement.  And with people like myself, and
90% of the average users, who simply prefer to use the best, that
ability to improve tends to weigh a lot more than the initial
innovation.

DOn't take me wrong, I'm not defending MS's business practices.  Simply
showing that where Microsoft wins, it might be a good idea to look at
and make use of the technology they used, because in many (if not most)
cases, it *was* better than the competition.

> 
> > If .NET doesn't take off, or doesn't have the impact I expect it to,
> > then yes, RUby probably will survive just fine as is, if perhaps near
> > non-existant on MS platforms.
> 
> According to the Ruby Garden polls, the combined Windows users account for 
> just over 34%, while Linux has 43%.  That's not bad saturation for Windows.

Yes, well, if .NET takes off, most of the Windows world will be running
on .NET - thus, Ruby would die off on that platform without .NET
support.

> 
> >> Hm.  Straw man argument.  Labeling my opinions FUD does not invalidate
> >> them.  I don't claim to be a subject expert.
> > 
> > Aye, that's mostly what FUD is, isn't it?  "Badmouthing" something with
> > no real technical merit.  ^,^
> 
> I define of FUD as misinformation disguised as fact.  Opinion is another 
> matter.  If Bill Gates says he hates Linux because it is hard to use, 
> that's opinion.  Paying a company to "produce" statistics that you define 
> and then offering that information as proof -- that's FUD.

Hmm, I suppose you could look at it that way.  I suppose that makes just
as much sense.  Sorry then for my earlier acusation.  ^,^

> 
> > what it can to make itself the leader of the technology - but opening
> > the standards, letting every major player make their own implementation
> > (both Sun and IBM have their own, for example, and Apple likely will
> > too) and then breaking that will have one heck of a backlash.
> 
> You don't see the difference?  Microsoft has "opened" the standards to MS 
> Word documents, too.  Anybody can implement their own MS Word format reader 
> and writer.  StarOffice even does a pretty good job.
> 
> But not good enough.  *That's* the point.

Likewise, MS doesn not read StarOffice format documents, even tho they
are open.  DOes that mean Sun is evil?

> 
> I'm going to re-arranging the order in which I respond to the rest of the 
> post.
> 
> > I don't see a Ruby# happening.. that would be fairly pointless.  Tying
> 
> Really?  They've had to do it to other languages that uses .NET, just to 
> get them to work.  J#, Eiffel#, VB.NET...

Extensions to the language only, as I read - no loss or change of
functionality.

> 
> > Besides, again, even if Mono can't communicate with MS *at all*, Mono
> > would still be one hell of a useful technology, one that Ruby could at
> > least make use of, if not greatly benefit from.
> ....
> > ^,^  Again, misinformation.  .NET does not in any way put your
> > information in MS's hands.  That is *only* if you use Passport or some
> > other MS service.  And there, Passport does not requrie any information
> > out of you for its basic usage.
> 
> Right.  Your argument was that we need Mono on Linux because Linux users 
> need access to .NET services.  Passport is a part of that.

Well, the one things that is pretty important to open use of .NET are
the alternatives to Passport.  XNS looks fairly good (I really hope
theyu open that soon, it looks like a great system), and there are
others.

If MS *does* pull something like not letting vendors use Passport if
they want to use another auth service as well, then... well, then I see
a big problem brewing with .NET.  One that does not make the core .NET
services unusable, just one that diminishes the usefulness of .NET of
Linux.

> 
> I'm still not sure what .NET is giving me.  I can already do IPC and 
> distributed computing across platforms and languages with XMLRPC.  What I 
> can't do is write agents in multiple language and have them run on any 
> other platform, which is what C# claims to offer -- although, with Java you 
> get a fairly good selection with JPython, JRuby, JavaScript, etc.

Java hasn't taken off all that well, tho, for multi-language stuff. 
Plus, again, some langauge simply can't be done efficiently in Java.

> 
> > As for C#... Well, I've not used it, but I'd certainly like to.  I like
> > C/C++, and the thought of a byte-code (with JIT) C++ like language that
> > is immune to things like memory leaks and buffer overruns sounds very
> > interesting.
> 
> IMO, nearly anything is an improvement on C/C++, although why anyone 
> expects C# runtimes to be any faster than Java runtimes is beyond me.  
> Java's had years to improve its speed and JIT compilers; MS is building 
> from scratch (we assume).  I haven't been very impressed with MS's 
> quality of coding, so I'm not expecting much.

The reason here is that C# was designed to be JIT, and thus some
language features are more easily compiled to machine code.  Even with
the best of the breed Java JIT compilers, the JIT itself takes a very
long time, causing a big delay in program start up.

> 
> > Well, that is your choice.  ^,^  I simply suppose we have different
> > criteria on what we base our usage of software on.
> 
> Yah.  I call mine "ethics".

Ah, I don't want to even get into that.  ~,^  My whole take on GNU/Free
Software is rather negative, so I'm going to avoid this argument.

> 
> Cheap shot.  But, really, when it comes down to it, that's what I believe 
> my criteria are.  I don't buy from companies that blatantly pollute and I 
> don't work with companies that exhibit a complete lack of business morals.
> 
> Anyway, you'll see Mono on Ruby, I have no doubt.  If I had any power to 
> stop it, I wouldn't.  I'm not RMS.  Ruby is in good hands, and I trust 
> Matz' judgement on where it goes.

Well, there we agree for sure.  ^,^

> 
> -- 
>  |..  "ROM stands for Read-Only Memory, meaning that new things cannot be
> <|>    written to it; ROM is the silicon equivalent of the brain of a
> /|\    religious fanatic."
> /|    -- Lincoln Spector
>  |