Re: Validating FO

On Wed, January 22, 2014 10:30 pm, Michael Hahn wrote:
> On 01/22/2014 07:19 AM, Tony Graham wrote:
>> On Tue, January 21, 2014 12:10 pm, Dave Pawson wrote:
>>> On 21 January 2014 10:25, Tony Graham <tgraham@mentea.net> wrote:
>> ...
>>>> In terms of validation outside the editor, thousands of people manage
>>>> to
>>>> run epubcheck [3] on their EPUB files to do a multi-stage check of
>>>> every
>>>> aspect of the EPUB.  If one schema can't express everything about FO
>>>> files, then perhaps users need a multi-stage checking mechanism rather
>>>> that whittling FO down to what can be expressed in a particular schema
>>>> language.
>>>
>>> Epub <> FO? Quite a different ball park.
>>
>> Running epubcheck --> happy EPUB producers and consumers
>>
>> Running 'focheck' === running epubcheck
>>
>> &#8756; running 'focheck' --> happy FO producers and consumers.
>
> To take Tony's observation one step further:

Yesterday wasn't my day to communicate well.  I was trying to both be
similarly equationish and to elaborate on the utility of a utility rather
than just drop the notion because EPUB is not FO.  I didn't think I was
making any new observations, so I don't think I was heading in the
direction your going.

> While I understand some folks are generating FO other than from XML
> through XSLT, that's my focus.  I want to know:
>
> 1) the XML input data is valid and correct*;
> 2) the transformation creates the FO document I expected; and
> 3) the resulting output has the correct appearance.
>
> Validate FO?  Don't care.  In fact, depending on the FO processor,

I think Dave also wants a schema as a guide for authoring.  I don't think
he's said it on this mailing list, but he's said in the past about writing
FO by hand to sort out what FOs and properties to use and then make his
XSLT generate the same.

> validating the FO can get in the way.  I explicitly have validation
> turned off for the FO engines (Antenna House, FOP, and XEP) I'm using
> since I mostly process draft (read: incomplete) input.  I'm more
> interested in having the processor do the best it can and give me an
> error log I can choose to ignore than have pristine FO as input to the
> processor.
>
> *As I told my students repeatedly, valid should not be confused with
> correct - DTDS and schemas can only go so far.  "Correct" markup falls

A schema is sometimes the best that you can get for help with guided
authoring, in the absence of custom programming.  Dave and I have
fundamentally differed in whether a schema can be enough.

> into the arena of an application and the human editor.  That's why
> things like authors' guides exist - "yes, the schema lets you do that
> three different ways, but we want you to do it this way..."

I think part of Dave's problem with the status quo is that between
inheritance and everything being allowed everywhere, there's rather more
than three places you can put things.

Regards,


Tony.

Received on Thursday, 23 January 2014 16:47:56 UTC