MathGroup Archive 2008

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: V.6.0.2 gripes...

  • To: mathgroup at
  • Subject: [mg86700] Re: V.6.0.2 gripes...
  • From: David Bailey <dave at>
  • Date: Wed, 19 Mar 2008 05:22:13 -0500 (EST)
  • References: <> <fras7l$1bn$> <frdfs4$avb$>

alexxx.magni at wrote:
> Murray Eisenberg ha scritto:
>> My experience with the Documentation Center in 6.0 and 6.0.1 is that the
>> first call to it is slow, but that after that things are faster at a
>> given session.
>> so could the problem be the initial loading of the Java interpreter, or
>> some other Java-related initialization, and not just initializing the
>> window per se?
>> Then of course there's the issue of how else to handle this.  One option
>> is to write everything in a platform-dependent way and avoid a
>> cross-platform tool such as Java entirely.  But that has its
>> down-side,too:  increased time to program it, higher cost,....
> I see of course the need to keep things simple when a project is
> dealing with multiple platforms - although I imagine that a
> development firm as big as Wolfram should be capable of handling it.
> Yet this argument doesnt convince me that Java was a good answer: what
> is more platform independent, for example, of a well written HTML
> documentation? Or, simply, the use of Mathematica notebooks without
> Java - as in the good old days?
> No, the huge slowness that many people experience (not everybody, I
> admit) is a faux pas.
> Java can be better than in the old days, as David Bailey says, but it
> still contains too many unresolved speed issues if 2GHz machines are
> crawling when simply reading documentation.
> What about an option to turn it off at least in the docs, losing some
> (unneeded) interactivity?
> Alessandro
>> alexxx.magni at wrote:
>>> It is just me, or the handling by Java of the new Function Navigator /
>>> Virtual Book is terrible? It needs a huge time to init the window, and
>>> its reponse time is awful, even worse than the Doc Center.
>>> It's an old and tiring camplain, but could people stop writing stuff
>>> in Java, since all they obtain is to slow our PCs to a crawl???
>>> Alessandro
>> --
>> Murray Eisenberg                     murray at
>> Mathematics & Statistics Dept.
>> Lederle Graduate Research Tower      phone 413 549-1020 (H)
>> University of Massachusetts                413 545-2859 (W)
>> 710 North Pleasant Street            fax   413 545-1801
>> Amherst, MA 01003-9305
I think we all agree that there should be absolutely no noticeable delay 
when bringing up a page of documentation - I hope WRI are listening!

Note that Java is not used to actually display the documentation - it 
appears in a notebook. HTML documentation would be greatly inferior 
however, because you could not actually execute documentation cells 
(possibly with changes).

I suspect WRI have bought in some Java search component that is too 
inefficient - hopefully they will re-code it in a faster form, possibly 
inside Mathematica, in C.

David Bailey

  • References:
  • Prev by Date: Re: Compiled function slowdown
  • Next by Date: Re: Using a logical Or in the function definition
  • Previous by thread: Re: V.6.0.2 gripes...
  • Next by thread: Re: V.6.0.2 gripes...