Re: Does the Crystal Goblet apply?

On 01/04/2014 08:47 AM, Tony Graham wrote:
> On Thu, January 2, 2014 3:45 pm, Arved Sandstrom wrote:
>> On a sidenote, Tony, I am personally a fan of SWIG:
>> http://www.nntp.perl.org/group/perl.macperl/2003/09/msg2874.html. :-)
> I first heard of SWIG from you back in your Perlish FOP days.
>
>> Dunno if SWIG is the answer, though, although it's something to
>> consider. At the same time I was associated with FOP I wrote an XSL-FO
>> formatter in Perl that captured quite a lot of the spec: that took less
>> than a calendar month. I obviously had the significant advantage of
>> knowing the spec very well, but I'd still say, based on that - and
>> general experience - that writing XSL-FO formatters from scratch for a
>> few more important languages could well be better than SWIG.
> So where did I go wrong with xmlroff?
Didn't know that you did, Tony. :-) Note that I didn't dismiss SWIG, 
hence I didn't dismiss SWIG operating on an xmlroff interface - I just 
said that SWIG may not be the [best] answer. I'm not always right, you 
understand.
> More to the point, who is going to write these multiple formatters?  Going
> to get them from 'quite a lot' to 'everything people expect to use'?
> Maintain them going forward?
In my opinion this won't actually happen, no. I'm simply stating that 
had XSL-FO been better advertised, I believe that programmers would have 
written more FO formatters in a wider variety of languages. It's a 
hypothetical observation.
>> I think you hit on a central point, which is education: tutorials, for
>> example. XSL-FO is not suffering low rates of adoption because it's more
>> difficult to use than other technologies, it's suffering because it
>> hasn't been sold that well.
> And if it's not officially part of the OWP and if the W3C is concerned
> about pushing the CSS 'brand', it won't be sold at all by the W3C.
>
> Writing tutorials and 'interface' libraries doesn't have the cachet of
> writing a new spec, but it may be more practical at this point.  However,
> I don't know how well it fits with your (I think) earlier point that
> developers don't grok XSLT to begin with or with Kai's point about ending
> up having to tweak the FO files by hand.
Yes, I made that point about XSLT. It's not meant to denigrate, it's 
just an observation that XSLT is quite complex, and not only do most 
programmers not use it, of those that do, not many use it enough to be 
proficient.
> Isn't this also the sort of thing you were arguing for a year or so ago or
> were you just on about conformance of existing processors?  If it was,
> forgive me for taking so long to wake up to your PoV.
>
> Regards,
>
>
> Tony.
>
Tony, over the years I've had many varied thoughts about XSL-FO. :-) No 
small number of these are due to frustration with slow and relatively 
minimal adoption of the technology.

Yes, I think I have argued previously for more education. Conformance of 
processors is also a point I have discussed.

Arved

Received on Saturday, 4 January 2014 20:48:57 UTC