RE: XMLSpy - FOP
-Original Message- From: Savino, Matt C [mailto:[EMAIL PROTECTED] Sent: February 28, 2002 3:20 PM To: '[EMAIL PROTECTED]' Subject: RE: XMLSpy - FOP Arved, I'd love to help out with the Perl prototyping if you have any pieces that make sense to break off. I hear you about the UML. I think on some projects it's more about control than functional necessity. [ SNIP ] -End Of Original Message- Anyone who has interest should track the mailing list for that project, which is available off the project page: https://sourceforge.net/projects/xslfo-proc/. I can think of some stuff already, which is best discussed there. If you need a break from Java and just want to get back into Perl or C, you're welcome. Also, this is an opportunity to come up to speed with XSL implementation while waiting for the FOP redesign to reach a less critical stage. If it's an either/or situation please stick with FOP and try to help out with the maintenance branch. Regards, AHS
RE: XMLSpy - FOP
Thanks Peter. I'm really glad to hear some positive news about the redesign! I'll try to get on that list and check out the code. I like reading Perl, it's therapeutic when you have to deal with Java all day long. I'm sure FOP is a very tough problem. I regretted the tone of that last email as soon as I spit it out. I'm just not having a very great day. I know from watching the fop-dev list for the last year or so that all the active players on FOP are very concientious, hard-working and know what they're doing. Maybe you can look at my little outburst as a rare window into the frustrations that some of the end users are feeling but have the tact to withhold. Just to give you an idea of my personal situation w/FOP, I work for a company of about 27,000 people. In the course of fighting to convince the higher-ups that we could do our little project just fine without Cognos or Crystal Reports, I've sort of become the chief FOP evangelist/programmer. (When you looked at our requirements and resources, building really did make more sense than buying--which is the whole promise of J2EE right?) We're trying to pull together a corporate culture that could have entire separate teams working on similar projects who didn't even know of the other's existstence. People are starting to look at my PDF solution. I'm just afraid they're going to look behind the curtain and see that I can't generate five 10-page reports at once (or one 100-page report) on one instance of Weblogic running on $80k worth of hardware--w/o running out of memory or coming to a standstill. I like where it sounds like you're going with the memory issues. I don't think the speed is a showstopper, but those out of memory issues sure are. I've upped my max-heap size to 768M. Anyone know of major pitfalls to this? (I have 1GB available per instance.) I guess even the roughest non-binding ETA on the redesign might help some of us sleep better at night - 6 mos? 1 year? 2 years? more? And just out of curiosity, why are you starving to finish this - love? future consulting gigs? both? Thanks again, that's all the rambling I have in me for now. Matt Savino Senior Systems Analyst Quest Diagnostics Clinical Trials -Original Message- From: Peter B. West [mailto:[EMAIL PROTECTED] Sent: Wednesday, February 27, 2002 4:43 PM To: [EMAIL PROTECTED] Subject: Re: XMLSpy - FOP Matt, I'm sure Arved will have something to say about this, but are you subscribed to SourceForge's xslfo-proc list? Have you seen Arved's perl prototyping code? This is a *difficult* problem. A large part of the spec can be implemented relatively easily, but if you get the design of the first, say, 85% wrong, the last 15% becomes well-nigh impossible. I *have* been living off savings and credit cards for over 12 months now, initially struggling with the original code base before deciding to start from scratch, and I can say that I am beginning to get a handle on the design. Good luck. Peter Savino, Matt C wrote: I'm 99% sure there is a huge corporate demand for an FO-PDF engine right NOW. Those guys at RenderX are nice but unresponsive, their product is on par with FOP at best, and the're too busy to breathe selling $5k/CPU licenses! I'm about ready to quit and starting working on one myself. Anyone who knows the PDF spec inside and out and can live off savings or credit cards for 6 months is welcome to join.
RE: XMLSpy - FOP
-Original Message- From: Savino, Matt C [mailto:[EMAIL PROTECTED] Sent: February 27, 2002 9:15 PM To: '[EMAIL PROTECTED]' Subject: RE: XMLSpy - FOP Thanks Peter. I'm really glad to hear some positive news about the redesign! I'll try to get on that list and check out the code. I like reading Perl, it's therapeutic when you have to deal with Java all day long. I'm sure FOP is a very tough problem. I regretted the tone of that last email as soon as I spit it out. I'm just not having a very great day. I know from watching the fop-dev list for the last year or so that all the active players on FOP are very concientious, hard-working and know what they're doing. Maybe you can look at my little outburst as a rare window into the frustrations that some of the end users are feeling but have the tact to withhold. [ SNIP ] -End Of Original Message- Hi, Matt Let me clarify. The redesign is what Keiron Karen (primarily) are working on. It is a redesign for FOP. What I started last fall is another project, which is intended to produce a C/C++ XSL-FO formatter. This is called xslfo-proc, and is on Sourceforge. I had a number of reasons for diverting my energies from FOP, and they had to do with limited free time last summer (due to intense real work) and a general burnout with how complicated FOP had gotten. xslfo-proc got off to a slow start, mainly because the company I was working for in real life went bust in October, and it's only this month that I got back into it. I did enough UML design last fall to realise that that was a waste of time. So a month ago I started working on a Perl prototype. I uploaded the first code yesterday but please don't expect this to actually be doing any layout yet. There is a lot there already, though, and at my current pace I expect to have some pretty good layout happening within a month. I have no intentions of abandoning FOP, but the redesign for FOP was and is critical, and only so many people can usefully do that - two max, IMO. When the redesign started I was out of the loop, so now I am waiting like everyone else. In the meantime I am devoting most of my efforts to xslfo-proc - it has a different approach and I hope it complements FOP rather than competes with it. My intention further down the road is to fold it back into Apache when the codebase is mature, there is a community built up around it, and the time is right. I'll second one specific comment of Peter's very strongly. I also believe that an XSL formatter project that is going to succeed has to tackle the whole problem. Formatting is not very modular, in other words. Not everything needs to be implemented right away but it sure needs to be considered, and a place for everything needs to be built in. The existing FOP shows us that retrofitting doesn't work. I'll welcome your interest in the xslfo-proc project if you feel like a break from Java. I don't mind admitting that since I work with Java every day in real life that the opportunity to get back to Perl (and C down the road) was not an unpleasant thought. :-) Regards, Arved
Re: XMLSpy - FOP
Arved et al, To clarify further: there are three re-design efforts going on. Keiron Karen in Java, building on the existing code base. Arved doing a ground-up redesign in Perl (protptyping) and C or C++, as he has discussed. Me, in Java, doing a ground-up. Do not despair. If Flannery O'Connor is to be believed, Everything That Rises Must Converge. I believe that these design efforts will, if not converge, at least cross-fertilise one another considerably. You will probably have noticed that there is a lot of cross-talk between the principals. At the end of the day, I think that the best ideas will shake together in the bottom of the pan. However, I would hope for more. I can conceive of no reason why a common design will not work in Java, C, C++, Perl or any other language of choice. Implementation details may differ, but the same overall design should be realisable in any useful language. In that sense, I disagree with Arved. I think his work *is* contributing to the redesign of FOP. Obviously, I think mine is too. So, if you decide to get involved in xslfo-proc, the effort will not be wasted in terms of helping bring FOP to completion. Let me just strongly endorse Arved's comment about the oxymoronic UML design, as in design by UML. What a bizarre idea. It must be something consultants do. Peter Arved Sandstrom wrote: Hi, Matt Let me clarify. The redesign is what Keiron Karen (primarily) are working on. It is a redesign for FOP. What I started last fall is another project, which is intended to produce a C/C++ XSL-FO formatter. This is called xslfo-proc, and is on Sourceforge. ... I did enough UML design last fall to realise that that was a waste of time. So a month ago I started working on a Perl prototype. I uploaded the first code yesterday ... I expect to have some pretty good layout happening within a month. I have no intentions of abandoning FOP, but the redesign for FOP was and is critical, and only so many people can usefully do that - two max, IMO. ... I am devoting most of my efforts to xslfo-proc - it has a different approach and I hope it complements FOP rather than competes with it. I'll second one specific comment of Peter's very strongly. I also believe that an XSL formatter project that is going to succeed has to tackle the whole problem. Formatting is not very modular, in other words. Not everything needs to be implemented right away but it sure needs to be considered, and a place for everything needs to be built in. The existing FOP shows us that retrofitting doesn't work.
RE: XMLSpy - FOP
So basically, the nuts are on the anvil? I hope no-one reads this mailing list. Or just run a search on Google in a few days. Better not put any fudges on your resume that can be contradicted by any post you've *ever* made to a newsgroup, mailing list or website. At least not if you have an uncommon name like mine. Matt Savino Senior Systems Analyst Quest Diagnostics Clinical Trials -Original Message- From: C Brian O'Kelley [mailto:[EMAIL PROTECTED] Sent: Wednesday, February 27, 2002 7:05 PM To: [EMAIL PROTECTED] Subject: Re: XMLSpy - FOP Random question, but is there incentive for a company to hire someone to enhance FOP and release the updates to the public domain? Are most of the developers employed to do this, or is this done in their free time? Brian On Wed, 27 February 2002, Peter B. West wrote: Savino, Matt C wrote: I regretted the tone of that last email as soon as I spit it out. I'm just not having a very great day. I know from watching the fop-dev list for the last year or so that all the active players on FOP are very concientious, hard-working and know what they're doing. Maybe you can look at my little outburst as a rare window into the frustrations that some of the end users are feeling but have the tact to withhold. I think your comments will be read in that light. ... I've sort of become the chief FOP evangelist/programmer. ... People are starting to look at my PDF solution. I'm just afraid they're going to look behind the curtain and see that I can't generate five 10-page reports at once (or one 100-page report) on one instance of Weblogic running on $80k worth of hardware--w/o running out of memory or coming to a standstill. So basically, the nuts are on the anvil? I hope no-one reads this mailing list. I guess even the roughest non-binding ETA on the redesign might help some of us sleep better at night - 6 mos? *** 1 year? *** My guess. I think the design is getting towards critical mass. See my other post responding to Arved. 2 years? more? And just out of curiosity, why are you starving to finish this - love? future consulting gigs? both? Things slowed down a lot after dot.con, and here in Brisbane they were slow to start with. I have a (voluntary) application for FOP, and I had a need to learn Java to spruce up my skill set. Future consulting? I suppose I can dream. Peter
Re: XMLSpy - FOP
Matt, Put this line at the top of your batch file: @echo 1:[ %1 ] 2: [ %2 ] 3: [ %3 ] 4: [ %4 ] 5: [ %5 ] 6: [ %6 ] 7: [ %7 ] 8: [ %8 ] 9: [ %9 ] t.out Then look at t.out to see what XMLSpy is sending to the batch file. Then try the same sequence from the command line to figure out what's going on and tweak the batch file to make it work. (That's how I made the batch file in the first place.) This is what I get: 1:[ -q ] 2: [ -xml ] 3: [ C:\Fop-0.20.1\tmp.xml ] 4: [ -xsl ] 5: [ C:\Fop-0.20.1\pdf_master.xsl ] 6: [ -pdf ] 7: [ C:\Fop-0.20.1\Output.pdf ] 8: [ ] 9: [ ] Chuck Paussa Savino, Matt C wrote: Thanks Chuck, I see I can actually use your workaround to go straight from XML-PDF with XMLSpy. But my problem right now is I can't even get XMLSpy to run FOP on an FO file that I know works. I just get that same error every time. (Regular XSL transformation works fine by the way.) I keep thinking there's a space in my file path or something, but I can't find any. I'll let you know when I figure it out. By the way I was working on this because I wanted to get your markers example going. I finally just did it on the command line. It looks pretty cool. I'm trying to figure out if I can use it to solve my adding '(Continued)' to table headers problem or my spearately numbered sub-section problem. Matt Savino Senior Systems Analyst Quest Diagnostics Clinical Trials -Original Message- From: Chuck Paussa [mailto:[EMAIL PROTECTED] Let's see Path to FOP bat file = C:\Fop-0.20.1\fopx.bat fopx.bat file = saxon -o test.fo %3 pdf_master.xsl @java -Xms256m -Xmx256m -cp build\fop.jar;lib\batik.jar;lib\xalan-2.0.0.jar;lib\xerces-1.2 .3.jar;lib\avalon-framework-4.0.jar;lib\logkit-1.0b4.jar;lib\j imi-1.0.jar org.apache.fop.apps.Fop -c conf/userconfig.xml %1 -fo test.fo %6 %7 %8 I guess I sort of got it to work but I had to go through an intermediate .fo file and I hard coded the xsl Chuck Savino, Matt C wrote: Has anyone gotten FO transformations to work on XMLSpy w/FOP .20.2? No matter what I try I keep getting the following error: Output of external XSL converter: The filename, directory name, or volume label syntax is incorrect. thx a lot, Matt Savino Senior Systems Analyst Quest Diagnostics Clinical Trials
