OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

oiic-formation-discuss message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [oiic-formation-discuss] The importance to users of documents looking the same

On 6/19/08, Simon Calderson <caldersons@yahoo.co.uk> wrote:
> ----- Original Message ----
> From: Peter Dolding <oiaohm@gmail.com
> Subject: Re: [oiic-formation-discuss] The importance to users of documents
> looking the same
>> > I'm not sure "most of the information" is correct. As a simple example,
>> > ODF styles can have information about
>> > how wide paragraphs ought to be, and how high the text in them should
>> > be, but says little about word width. With
>> > that in mind, words could be different widths, causing different line
>> > breaks and therefore different pagination. It's
>> > difficult to see how you could do deeper testing than what vendors are
>> > already going to be doing anyway, and from
>> > that point of view it doesn't seem to make sense to ask a TC to do that
>> > work also.
>> It makes perfect sence for TC to do it.  This way all applications only
>> need to keep one set of test systems.  On top
>> its exactly the same test system not something a coder somewhere has made
>> a error creating and not had it double
>> checked leading to a defective ODF supporting applications..
> It makes sense for the new TC to develop test documents. It doesn't make
> sense for it to test or to feed back on issues/goals which are outside the
> scope of the ODF TC. I don't know how you can aim for pixel-perfect layout
> without compromising on separation of content and layout (which is a
> specific TC goal).
What you are not getting I am not outside the edge of TC coverage.

Sorry to say "compromising on separation of content and layout"  ODF
rendered layout is dependant on content processing.  If it was not you
would not have images and the like connected to a particular line of
text.   So that has to be a TC error when covering rendering of the

If you want separation between content and layout in render please go
to PDF and all the document editing problems that causes.

I more use the term Layout Perfect.  I will tollerate this may not be
do able with 1.1 ODF.   The information in ODF allows for applications
to be tested for layout/rendering inside the rules of ODF..

I am after clarifications for the simple reason I can build it now.
But I am looking at some really large tolerance.   So bad that over 20
pages one program could be 20 pages and another be 21 and still be
inside tolerance.    Even worse that difference could the what you see
is what you get.  Ie application on screen telling you 21 pages when
you print you get 20 pages.  That is quite a nasty one.

Now rules on layout maths and font metrics scaling.   Can cut my
tolerances right down to + - 2 This is not pixel perfect this is still
classed as layout perfect.  You can depend on layout being side
tolerance so document will not ever have anything more than minor
errors.   Since a pixel could be anywhere inside 4x4 box and be
correct.  Compared to by some points threw document now without
clarifications being to + - 1 000 000 or higher on the document  So 2
000 000 x 2 000 000 as long as the point is somewhere in there its
correct.  Key thing  2 000 000 x 2 000 000 is larger than most peoples
computer screen res.

Now to avoid such a completely useless tolerance level  it will cause
me to build up  8 different tolerance settings to cover all the nice
grey factors.  So every application will have to have its rendering
system tested against 8 different sets of maths.   And as long it
passes one it passes.  That is vial of course the tests are still
inside what information ODF standard provides.

I am simply after some answers.  With answers test makers can develop
tests of ODF application rendering.  This will for ever draw a line in
the sand.  Past this point your application is now not rendering
layout as per the information the ODF file provide.  You have to have
a math error somewhere.  Of course I want as low as tolerance levels
as I can in the test.  It makes test writers job simpler.  I am
focusing on visual rendering.   Others can focus on audio rendering
for the deaf.

Peter Dolding

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]