XSL-FO support in Publishing Engine 7.0

[ClayHelberg]

Hi Folks--

I know that in previous versions of PE (<=6.x), XSL-FO support was implemented in a multi-stage approach, where the document is transformed into XSL-FO markup, and then a dynamically generated FOSI is used to create the final output. Can anyone verify whether that is still the case in PE 7.0? There was talk a while back of developing XSL-FO support via APP rather than via FOSI, but I haven't heard whether that actually went anywhere.

Thanks!

--Clay

[GarethOakes · Accepted solution]

It still works just like it ever did, there has not been any significant new development to XSL-FO support in Arbortext since Paul Grosso left. As you say: XML --(XSLT)--> XSL-FO --(ACL)--> FOSI --(TeX)--> DVI --(dvi2pdf)--> PDF.

I guess it would be technically possible to implement a subset of XSL-FO on top of APP but that hasn't happened and it's not clear if XSL-FO support is currently a focus for PTC.

[ClayHelberg]

Thanks, Gareth. I suspected as much, but wanted to confirm it.

Just out of curiosity, why do you say "subset" when you mention the possibility of implementing XSL-FO on top of APP? I would have thought that APP's capabilities would be a superset of XSL-FO, so if you took the time you ought to be able to do a pretty complete implementation. Is there something specific you're thinking of in XSL-FO that would be tricky to layer on top of APP?

--Clay

[GarethOakes]

Regards page layout and typesetting horsepower, sure, APP is able to build more complex layouts and solve more complicated problems than XSL-FO. What I meant by the word "subset" is related to the XSL-FO specification, which clearly defines things like: typesetting behaviour, whitespace collapsing behaviour, block model, and so on. There are variations in the algorithms and approach of APP composition versus the XSL-FO specification. FOSI is actually more closely aligned with how XSL-FO "thinks" about page layout and typesetting (I suspect the same or similar people worked on both specs).

Last time I looked closely at implementing XSL-FO in APP there were enough differences in behaviour that it would be very difficult (costly) to correctly or completely implement XSL-FO on top of APP. APP now has a more robust block model so I'd guess it would be a lot easier these days. Either way I think it would still be a bit of a struggle to achieve 100% compatibility with the XSL-FO spec.

[ClayHelberg]

OK, that makes sense. Thanks for the additional explanation, Gareth.

[ClayHelberg]

Hey Gareth--

I was just surfing around on the web for more information on XSL-FO engines, and I came across this:

The Cover Pages: Advent 3B2 Releases 3B2-FO XSL-FO Rendering Tool

I wonder what ever happened to this.

--Clay

[ChrisWestern]

Hi Clay,

3B2-FO was quite quickly dropped by Advent when it became clear back in 2003 that 3B2 would not be able to support the full FO spec in the way the standard required (e.g. block model as Gareth mentioned) promtly followed by its partner product 'pure' a simplified seudo FO type engine.

In the mid/late 2000's a partner of ours implimented a FO subset solution with 3B2 V9 for a large cutomer in France with some success so i believe.

With V11.1 however i expect an FO library would cirtainly be possible and something we have discussed developing with PTC, but with other much more competitively priced seperate FO engines around i expect this initiative is unlikely to get of the ground any time soon.

Best Regards

Chris

[ClayHelberg]

Thanks, Chris!

[GarethOakes]

Pure! That sure brings back the memories. Amazingly I found a few links still out there!

http://www.topologi.com/products/tmentor/index.php

http://esd.element5.com/product.html?productid=529824

http://www.bloomberg.com/research/stocks/private/snapshot.asp?privcapId=8841094

[GarethOakes]

Those were the good old days! *sniff* Chris covered the situation pretty well, it is an interesting historical footnote; from memory at that time XSL-FO was considered a threat to the lower-end 3B2 solutions so it seemed logical to "get with the program".

If you write an FO library for 3B2, let us know

[ChrisWestern]

I had the pleasure of presenting that little gem to the 3B2 users group at an XML Europe event, i think no one else wanted to try!

"Ahh" PPSL we were going to change the world back then, how naive

[RayStachowiak]

Hi Clay and Gareth and others

In 7.0 F000 release notes they included the following:

FOSI and XSL-FO Print Engines in Sustained Support

The FOSI and XSL-FO print engines have been placed on sustained support with

this release. They will not receive enhancements or maintenance fixes in the

future.

APP is the recommended engine for print output.

I took to mean no more updates and no more continuation of FOSI or XSL-FO. However, There was a serious bug that crept in in 6.1M040 that would not run XSL-FO They fixed this issue in 6.1-M080 and 7.0 M020. I know since we had and reported this issue from our Military Group. But it does not seem PTC is planning of adding XSLFO to APP.

[ClayHelberg]

Thanks for the additional info, Raymond. There would certainly be advantages to having an actively maintained XSL-FO publishing path in PE, but I can also see why PTC would decide from a business perspective that it's probably not worth the cost of developing it.

[AllissaLeach]

We here at Redstone Arsenal use XSL-FO extensively. We have reported an issue with the XSL-FO crashing (CASE ) which we did fix the XSL-FO.  NOw we have the same issue, but we have been told that FOSI and XSL-FO are under sustainment.  Whatever the issue that was fixed previously it is still broken.  PTC may have thought they fixed it, but the bug(s) are still lurking out there.

We are using 7.0 M030.  Right now we have several thousand page TMs that we cannot publish.

[RayStachowiak]

What issue are you having, error. The error that was fixed in 7.0M20 and also in 6.1M80 that we had revolved around the composer not pulling in the required FOSI along with the XSLFO styles. It needs the screen FOSI to set up some values I was told in addition to the XSLFO styles. This was fixed in 6.1M0 and I believe also in 7.0M20 . Our Military work packages publish fine.

[GarethOakes]

It's a shame but PTC has put XSL-FO out to pasture. If PTC support cannot help, there are partner companies such as ours (www.gpsl.co) who may be able to assist, as we have the skills to debug and patch the XSL-FO implementation in Arbortext (ACL and FOSI). Otherwise you could look at alternative XSL-FO software, but in that case keep in mind your stylesheets will need some work and also none of the other tools integrate directly with Arbortext. Again there are partner companies who could help with that approach.