[geeks] Browser licensing?

[Alois Hammer]

On Thu, 4 Sep 2008 01:58:55 -0400, "Joshua Boyd" <jdboyd at jdboyd.net>
said:
> 
> The actual startup speed of JVM or .NET have no relevent here as they
> clearly aren't trying hard to start quickly, nor do they show you the
> entire possible range of ways for a VM to start.

I was referring to the instantiation of J2SE either as standalone or as
a Firefox plugin on this machine.  Startup time is essentially
identical, even though we'd certainly like the browser plugin to start
as fast as possible.  Unfortunately, the plugin has to drag the rest of
J2SE with it.  I assume that Sun has finally realized that
Java-on-the-server is about as far as Java is likely to go for the
foreseeable future, but I remember a long marketing slog where Sun
wanted everyone running their apps off the Web and/or network, quite
possibly through a browser.  It's reasonable to assume that they spent a
fairly long time making the plugin as fast as possible until giving up--
if they have given up, which I doubt.  They're still trying to make
everyone use network terminals, and they've only been pushing that on
and off for twenty or so years.

And, since Sun spent a very long time trying to convince people to use
Java for desktop apps (and a lot of companies fell for it), it's also
reasonable to assume that the non-plugin case should involve a fair
amount of time spent trying to make the JVM not-a-hindrance, but darned
if OOo doesn't still take bleeding forever to start with Java support
turned off, and quite a lot longer with Java on.  And golly, look,
they're both primarily coded by the same company, which should be able
to pool its knowledge to make the whole package move faster as a piece.

FWLIW, Microsoft's release of .NET 3.5SP1 (which includes 2.0SP2 and
3.0SP2) includes quite a lot of work trying to make it start faster, so
Microsoft has apparently recognized what everyone else has: writing to
.NET means trading off whacking great amounts of performance.  Google
".NET Client Profile" or see
http://blogs.msdn.com/bclteam/archive/2008/05/21/net-framework-client-profile-justin-van-patten.aspx

And, yes, it's true that an 0.2 release of Chrome probably isn't 100%
indicative of what the final release will deliver.

On the other hand?  I believe that Google Search is, if not the only,
then nearly the only service or software they offer that doesn't still
have a "beta" tag on it, and with good reason.  GMail's only been in
beta since the first invite codes went out years ago, and it still
periodically eats your mail and burps.  Gears is still down for long
stretches, Maps results are getting damned unreliable for what *should*
be a pretty good case (Washington D.C. Metro area), and I can spend two
or three hours at a time being knocked off Google Search with a notice
that I look too much like a spambot today.  So I'm pretty confident that
if Chrome sucks now, it won't suck much less later.  I'd point to Picasa
as an example of good work from Google, except that it's more properly
Idealabs' software; not Google's.  And if I'm manipulating photos on
Windows, I use FastStone because it's, um, faster.  And does more.  And
crashes less often (never).  Even the beta releases.