Lists Home |
Date Index |
performance part 1. using xlstproc (xmlsoft.org)
stylesheet approx 50k
xml in approximate doublings in size from 24k to 3.25M (7 test files)
the processing time goes up by a factor approx 4 for every doubling in
size - looks o(n2) to me.
now comes the hard work, because i didn't write the algorithms.....
Tatu Saloranta wrote:
>--- Rick Jelliffe <email@example.com> wrote:
>>That is why, ultimately, the current XML efficiency
>>problem is not with
>>technology: not APIs, algorithms, CPU instruction
>>sets. The XML
>>efficiency problem is with motivating researchers
>>and open source
>>developers into areas that match corporate business
>With all due respect, this is undermining existing
>efforts that do try to address various problems.
>Many of people discussing approaches are actively
>working on various parts of the problem; although I'm
>most familiar with Wolfgang's and Michael's work, I am
>sure others are as well. And at least these two seem
>to be doing quite well in their respective fields.
>Also, even if we disagree to a degree on relative
>merits of various points of optimizations, I doubt any
>one of us is convinced to stop working on our
>'favourite' areas, perceived bottlenecks. If someone
>wants to focus on perfecting byte-by-byte level
>close-to-hardware utf-8 - to - xml - tokens
>transcoding, great! Once results are in, we'll see the
>benefits. If outsiders think this is a stumbling block
>then so be it: I am less worried about perception than
>actual motion towards the goal. But then again, xml is
>just a hobby for me, not my bread and butter.
>Finally, not all progress is done in public by
>(academic) researchers, either. If and when there are
>business opportunities (which seems to be the case for
>efficient xml message processing -- IBM is making
>killing selling their expensive hardware, n'est pas?),
>there is and will be progress. Value of prices like
>X-prices, or D-Arpa's vehicle challenge, is lately
>hyped to stratosphere. It will take a while to
>objectively evaluate true value. So while here, too, I
>think it would be interesting to see some prices, I do
>not see lack of such prices as a block that prevents
>progress. But if such prices come available, great;
>the more approaches the merrier!
>-+ Tatu +-
>Do You Yahoo!?
>Tired of spam? Yahoo! Mail has the best spam protection around
>The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
>initiative of OASIS <http://www.oasis-open.org>
>The list archives are at http://lists.xml.org/archives/xml-dev/
>To subscribe or unsubscribe from this list use the subscription
tel;cell:+61 411 287 530