July 16, 2004
I've spent the last couple of days with Tulna, building PDF reports with Inetseft's StyleReport Pro. It's a piece of shit. My brain is numb.
Posted to Java by Simon Brunning at July 16, 2004 04:47 PM
I've had a laugh actually...although it has been a nightmare, I've been pi**ing myself silly listening/watching SB disintegrate ;-)
If Simon had his way he'd spend 8 weeks building some powerful PDF generator, and then spend another 8 weeks trying to get something decent looking out of it.
The bitching is normal anyway, at least he spent his adding value rather than tinkering under the bonnet.
Really? You should have tried the other tools I QA'd for the job, really guys, that was easily the best one! What's the problem guys?
I liked it and thought it was easy to use, I dunno, you can't get the staff these days....
I wouldn't write it from scratch, Paul. I'd use iText. Or, even better, ReportLab.
With a good high level tool, the easy stuff is easy. The hard stuff might be impossible, but that's the price you pay. With a lower level tool, as I'd prefer, the hard stuff is possible, but you might have to do a bit more work on the basics.
With StyleReport, the easy stuff was hard, and the hard stuff impossible. Even the medium stuff, though possible, was a bit of a bugger. :-(
I'd use RootRiver Delta (http://www.rrsys.com). Works fine and makes even the hardest stuff possible, as far as I'm concerned.
I love LaTeX. A Python script spewing out a LaTeX file and then running pdflatex works well, but adding graphs is a pain, though possible with something like gnuplot.
I used to use iText, now adays I'm putting the data for my document in xml and using an xslt template to convert it to xsl:fo and putting that through fop to generate the document. I'm using XMLSpy to generate the stylesheet. Writing a report is now as fast a writing a word document. I feed xmlspy the schema for the xml and I can drop live values into the document I've written. Implementing this once I'd figured out the various parts where out there took under a day. Now if only there was a free/open alternative to XMLSpys stylesheet designer, then I wuold be set.
It is probably getting a bad rep here because of the bugs in the earlier version. With the newer 6.5 and 7.0, it has improved a lot in that department, and is at least as stable as the other tools I have used.
It's amazing how a blog, for whatever reason ranked high on google, is used to deface a fine product. I have used stylereport and it's fine for the intended usage it's designed for. If you use a hammer to drive a screw, it will be a bad tool no matter how good the hammer is. My impression is simon prefers to write code to generate reports instead of deal with point and click, which is fine but that doesn't make the point and click tool bad.
I'd have been happy enough with a point and click tool if the bloody thing hadn't been such a complete and utter pain in the arse to use. And this wasn't just my opinion - everyone I know who has tried using the product for real work hates it too.
I small subset of the problems I had:
* The WYSIWYG view, the preview and the printed report all look different.
* You can't have bold text in a justified paragraph.
* The design tool was a memory hog, and unstable.
* It's sometimes impossible to select the required component of the report other than going in via the object tree - which completely closes down after each item is selected.
* Moving or resizing report items often arbitrarily moved other items, creating a cascade effect, buggering up your whole report.
* 'Cutting" a report item sometimes deletes it.
* Inflexible report break options.
* More here: http://www.brunningonline.net/simon/blog/archives/001543.html
The support team were not interested in having even the most blatant of bugs reported to them without a very expensive support contract.
This may all have been fixed in later versions of the tool - but the version we used was bad enough that we aren't interested in looking at new versions.
BTW, I did like "for whatever reason ranked high on google". Had the English been less appalling, I might have been suspicious that this was someone trying a bit of spin. But if it was someone from PR or marketing at InetSoft, they'd never have been so stupid as to say "deface" when they mean "defame".
Simon, I wasn't accusing you of anything. You had your reasons and I just tried to offer a different viewpoint. With your superb command of English, I'm sure you could have tried to make your points without resorting to name calling.
If you wern't intending to question my credibility, Abram, you should have been a great deal more careful about the language you used, since that's clearly how it read.