Thursday, November 3, 2005 Paul wrote:

> Before I start - I love OOo and much prefer to use it than M$ products
> - I'll continue use OOo and would prefer to use OOo than pay M$.
> However, I think we need to get real and look at what is what...

I don't think this premise is necessary :). As it was
repeatedly mentioned, is not "talking about OOo" the
problem, is the style and content.

> === OOo slowness is nothing to do with parsing XML ===. If the
> article is re-read, it clearly shows that there was a test done for
> Excel opening an .xml file and OOo opening a .sxc file.

> Both products had to open XML files. I took both of these files and
> did some little tests of my own. The .xml file was opened in wordpad
> previous to the tests to ensure it was 'true' xml

> Excel opened .xml file in about 2 minutes
> OOo opened .sxc in about 4 minutes
> OOo opened the .ods version of the file in about 3 minutes

> Therefore both programs were having to parse XML, OOo was considerably
> slower than Excel. OOo was slower than Excel - its a fact and I don't
> believe it has anything to do with parsing XML - since both programs
> had to parse XML to produce results.

More about this later.

> I also believe it has nothing to do with compression. Again I tested
> compressing the .xml file and it took ~ 30 seconds. Even adding that
> to the times, the difference is lessened, but not completely closed.

Yeah, I was already called on that one :) I had done the
mistake of launching the decompression while OOo was opening
up, so the CPU was already hogged by OOo :)

> So it is not compression and it is not XML parsing. This means that it
> is how the programs handle the XML parsing and the internal code
> optimisation or something else.

Oh, I never meant to imply that the slowness was because of
the idea itself of XML parsing! I was obviously referring to
the way OOo was doing it. A part that obviously needs a lot
of work.

> Whilst these tests were not scientific both the article and my cobbled
> together tests indicate that it is OOo code that needs to be improved
> to make it faster. Its that simple.

Nobody is arguing with that.

> On this point I agree with Chad.. Lets call a shovel a shovel and get
> on with getting it corrected. I believe that the OOo developers would
> have this in hand, which is the best place for it to be...

Chad did not say to call a shovel a shovel and get on with
getting it corrected. He just starting spouting nonsense and
insults about how stupid the idea of having ODF as a
standard format was, when not even the reference program
could 'get it right' (time- and memory-wise). See the
difference, and why emails like the one you posted don't
need 'excuses' header? :)

-- 
Giuseppe "Oblomov" Bilotta


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to