Re: [oiic-formation-discuss] (1)(f) and (1)(g) -- audience and working language

["Dave Pawson" <dave.pawson@gmail.com>]

2008/6/13 Peter Dolding <oiaohm@gmail.com>:
> On Fri, Jun 13, 2008 at 7:21 PM, Dave Pawson <dave.pawson@gmail.com> wrote:
>> 2008/6/13 Peter Dolding <oiaohm@gmail.com>:
>> Unless that 2% pushes a text flow over onto the next page,
>> leaving a blank page in the chapter where there wasn't one.
>>
>>
> That is a layout failure.  2 percent + - is with layout intact.   If
> any text goes onto a different page what would be a different relative
> location it a failure to maintain layout.  So not layout perfect.

You offered 2% as a viable limit I think Peter? Zero tolerance?
Is that needed? Manageable? Font discrepancies could provide
some variance I think?

>
> Layout perfect is exactly that.  If a page has X number of lines on it
> and that is the way the person intended it that is the way it stays.
> No matter what happens to the page size.  Reason size of font and
> distance lines are apart is relative to printable area of page.

And what about images? text styling etc. All could provide some
variance. We are not talking about a page description language
where your points are perfectly valid.


>
> Everything must stay at its relative location and its relative size to
> everything else on the page to be layout perfect.  Most commonly done
> by relative size and location to printable area of page.

Yet we don't have any reference? Author intent is no good.
We have no reference implementation, ODF isn't page based is it?


>
> That is Layout perfect.  Its not something you can straight up use a
> normal ruler.  Of course if test document prints its own ruler you
> could use that.  Since that would have to be to the same scale.

I still can't imagine an objective test based on that Peter.


>
> Layout perfect but wrong size is something you do have people say in
> printshops.    That is exactly that the customer either want the
> document smaller or larger scaled without changing the layout in any
> way.

I'm not happy bringing scaling into a test, since I can't find
any trace back to the standard. I can see your logic for
using it when printing, but your reasoning for not
accepting print output seems logical.


>
> Issue is not many programs operate in a layout perfect way.  Lots of
> people don't have a clue that its even doable and compare the
> experiences to the layout imperfect tools they know.
>
> Dave Pawson you appear to never worked with software that generates
> layout perfect.

Agreed. XSL-FO is the nearest I have come to typesetting at all.


>
> One disadvantage in layout perfect can be page wasteful since it will
> not pack more text on a page.   The important but here is when
> printing a manual or contracts and the like where you want everything
> on the same page no matter what its a really nice feature because turn
> to page 10 on all documents created in a layout perfect way is exactly
> the same page just different scaling no matter what size page its
> printed on.

Agreed. Which is why lawyers like PDF :-)
The easy option is to waste space!


>
> Also layout perfect throws lots of ideas of font sizes images sizes
> out the window and goes more relative.  More real to producing exactly
> the same page or the closest representation the output system allows.
>
> Really Layout Perfect a form of imperfection tolerance.

If it is relative, please tell me what it is relative to;
Then provide an example test, that might be automated,
and that would provide two vendors with a warm feeling
about interoperability.


>
> We all know the number of lines of ink a printer does down and across
> the page is different between models on top of that there is
> differences on printable area on the same paper size.  Then we have
> different page sizes on top of that.  Yet for some reason lots of word
> processing programs got the idea that solid set sizes could some how
> work.  Simple fact they don't.  Text moves between pages people end
> being caught out by a person using 100 percent of page and there
> printer only able to print to 99 percent of the page area.  Even worse
> person does not have right size paper.  Layout Perfect due to allowing
> scaling these problems are a non issue.  Person can still print the
> doc still get it in a usable form.  And still be able to use that doc
> of a different scale effectively with everyone else.


All good reasons not to use 'layout perfect' as a viable test approach
for 'rendering' tests, as Michael calls it.

Do you agree with that?


>
> If the text ends up really small there can be issues with Layout
> Perfect.  The layout is still perfect at that point but the document
> is unreadable.


Hey, wait till your eyesight is as good as mine. "Really small" starts
to become 12point.
Seems to happen with tens of years in front of a computer.


> There is simply no such thing as a pixel perfect print.   But there
> are such things are Layout Perfect prints it is the closest thing to
> pixel perfect prints that can be created on every printer out there.

But is it usable, by this group, for the IIC TC, for testing for compliance
at the document or application level?

regards




-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk