discuss-gnustep
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: System fonts


From: Kazunobu Kuriyama
Subject: Re: System fonts
Date: Sat, 28 Aug 2004 23:44:25 +0900
User-agent: Mozilla/5.0 (X11; U; Linux i686; ja-JP; rv:1.4) Gecko/20030624 Netscape/7.1

Yen-Ju Chen wrote:

>> From: Kazunobu Kuriyama <address@hidden>
>> To: address@hidden
>> CC: Yen-Ju Chen <address@hidden>
>> Subject: Re: System fonts
>> Date: Sat, 28 Aug 2004 15:34:48 +0900
>>
>> Kazunobu Kuriyama wrote:
>>
>> >
>> >You often raise an issue on performance of text rendering. So I'd like
>> >to make
>> >sure of this: how much is the performance actually degraded when
>> Chinese
>> >text is
>> >rendered in comparison with pure ASCII one?
>> >
>> I don't need a detailed or scientific data. Rather I'd like
>> to hear your impression on that as a user.
>
>
> i can do some tests, but my impression is that if I load a Chinese
> text file
> with about thousands words in Ink.app, it takes 1-2 seconds to display
> it.
> I see the blank text view, then text comes up later. It's noticable.
> There is no problem in later editing, just the first time to load it.
> Actually the speed of GNUstep doesn't degrade in Chinese environment,
> since user interface usually don't have many text.
> It only happens with large amount of text.
> And as far as I remembered, iconv (which is used in GNUstep) is not
> very fast.
> Therefore, it may be caused by iconv instead of the implementation of
> GNUstep.
> I expected font substitute will slow thing down because it has to
> check the fonts frequently.
> But if it is slow doen from .01 second to .03 second,
> it is slower, but no problem for users. That's what I really mean.


Let me ask some more:
(1) What happens with X applications other than GNUstep? Does they work
as slowly
as GNUstep apps?
(2) How about a very huge ASCII text file? Does it also get GNUstep apps
to crawl?

Since any conjecture on performance can be quite harmful, I'd very
appreciate
it if you could identify the real performance bottleneck

<snip>

>
> As long as the first implementation comes out,
> we can have better idea about the pro and con of font substitute.

Since the font substitute mechanism is a widely used, established
technology, I think it's useless to discuss pros and cons here unless
we're seeking for a new technology better than that. Are we?

All materials we need are already put before us for free for years. Why
does GNUstep hesitate at taking them all for its own sake? Having font
substitite is no longer a challenging problem.

Thanks,
- Kazunobu Kuriyama






reply via email to

[Prev in Thread] Current Thread [Next in Thread]