Re: [Silgraphite-devel] limit on segment length

[Roy D. <roy...@ds...>]

Hi folks
As a newbie, I might have got it wrong but at the opposite end to the
"massive paragraph" usage case could be very small segments caused by
frequent changes of font for stylistic or linguistic reasons. Would
these case make the segment header more significant? Could this have
been the test case for the FW report? 

Change of font may not be perceived as a costly activity and I could
envisage segments for individual words. For OpenOffice spreadsheets,
columns with different font, or maybe even without change of font,
segment lengths could be short.

But there is also mention of the number of segments causing a crash in
FW. Are we clear whether the issue is number, size or total size of all
the segments? 

Sharon's calculation 100 * 342 yields 34k rather than 3.4k I think.

If a page is approx 5000 (80*60) glyphs does that map to 1.7MB segment
memory per page, or 340MB for a 200 page document? 

(My turn to make an arithmetic error no doubt!)
What difference would number( or average size) of segments make to that
(virtual) memory requirement?

Also see below.
regards
Roy Dalpra

On Wed, 2007-10-03 at 12:57 +1100, Joh...@si... wrote:
> 
> 
> > 
> > By changing these to 16-bit values we would start to impose a limit 
> > of something on the order of 32K of these in a segment. Of coures 
> > that would be way more than plenty for a real segment that is being 
> > rendered on the screen. But consider the sort of segment that 
> > incorporates an entire paragraph that is created for the purpose of 
> > doing high-level layout--if you had 200 lines of text and 150 
> > characters/glyphs per line, you'd get up near the limit. Now that is
> > a pretty huge paragraph, but maybe not outside of the realm of
> possibility. 
> > 
> > Do you all think this is a reasonable limit to impose in order to 
> > save some space? We'd probably need to introduce bounds checking in 
> > places we don't bother right now. :-( 
> 
> On the whole, I think not. For a 5% saving, you're imposing a
> requirement on all clients, with no (obvious?) way to work around it
> if their pesky users make a giant paragraph. 
> 
> For example, I've been known to make a single paragraph out of an
> entire book of Scripture (for a particular kind of Bible study). That
> could go well over 32K.

>  
But is a limit on the segment length really a limit on paragraph size?
Surely not. 
> 
> JohnT
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2005.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________ Silgraphite-devel mailing list Sil...@li... https://lists.sourceforge.net/lists/listinfo/silgraphite-devel