xmlgraphics-fop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Glen Mazza <grm7...@yahoo.com>
Subject Re: Valid version of the PDF specification?
Date Mon, 11 Oct 2004 16:58:01 GMT

It's not just the *quantity* of consumers, it is also
their level of happiness, as well as our ability to
recruit developers willing to do the coding.  (Absent
committers willing to spend the time on the code, *no*
version of PDF will get supported.)  Per your
suggestion, though, if we were to drop the FOP's PDF
specification-level to pre-2000 versions, these
problems occur:

1.) The 98% that would have no problem with the 2001
spec will note that our product doesn't render as well
as the commercial products do.  Features they would
want to implement aren't possible.  So the benefit of
making the 2% AR4.0 user base happy comes at a price
of making unhappy the other 98%, and due the ratios
involved, overall satisfaction with the FOP product
would fall through the floor.

2.) It is next to impossible to recruit people to
study the PDF specification inside and out in order to
implement a PDF renderer--there are no shortage of
other much-more-in-use skills that take far less time
to learn.  That problem is compounded when you ask
them (1) to not only study the PDF spec, but a
commercially obsolete 6-year old version of it, and
(2) to spend their time doing so with the full
knowledge of when they're finished, everyone is going
to be panning their work (see #1 above).  Few
developers can afford to do what you are proposing.

3.) Generally speaking, virtually no one is stuck at
AR4.0, and those who are aren't writing FO documents. 
You may make that calculation that it is acceptable
for 500,000 users to have a substandard product just
so the feelings of 10 AR4.0 users don't get hurt, but
I don't think you'll get much support for that on the

4.) By having a substandard PDF renderer, you lose
more V5/V6 people than you gain of V4 people, so
overall application usage drops rather than increases.
 Without exception the V5/V6 people are the ones you
would want to please anyway--they're the ones writing
the reviews, they're the ones more likely to be using
it for production applications.  (Someone from IBM or
Sun isn't going to write a review of "Using FOP with
AR4.0"; instead they will write about "FOP with
AR6.0", and in particular how poorly it performs and
how out-of-date it is.)

Glen (Who thinks AR V4 will work with the 1.4 spec
anyway ;)

--- Clay Leeds <cleeds@medata.com> wrote:

> On Oct 11, 2004, at 7:06 AM, Glen Mazza wrote:
> > 1.4 (Copyright 2001) is fine for us.
> That's OK by me, as long as that is the consensus
> and not the opinion  
> of one of our esteemed committers. I was hoping for
> a little discussion  
> about why PDF 1.3 is not even an option? There is
> some discussion of  
> this topic in the archives[1] & [2] and also in the
> FOPProjectTasks  
> wiki[3]. I acknowledge certain items require PDF 1.4
> (transparency,  
> encryption, and others[4]). But like Jeremias said
> in this thread[2]:

View raw message