[pro] Professional Reports

[Vladimir Sedach]

I just remembered one project I came across recently:

http://code.google.com/p/wkhtmltopdf/

I would consider using that also, and wish I had come across it
earlier. Templating or generating HTML documents is very convenient,
and if you're doing a web front end, you have all the libraries you
need already.

Vladimir

On Tue, May 24, 2011 at 3:30 PM, Vladimir Sedach <vsedach at gmail.com> wrote:
>> if i were to start today, i'd either generate libreoffice's xml format
>> (you can exec commandline tools to generate pdf);
>
> I tried this two years ago and ran into some problems. As I remember:
>
> OpenOffice (of which Libre Office is a fork) had a weird client-server
> model for command-line use. It was a pain to set up initially, and the
> version of OpenOffice that came with Debian on our servers didn't work
> at all. The best way to check if a document would finish being printed
> as a PDF was to poll the filesystem. It took a long time to generate
> the PDFs, and what's worse is that this time varied widely with no
> apparent causes. The whole process was unreliable.
>
> The templating itself was tricky. As you mentioned the node order
> matters, but also there were all sorts of undocumented dependencies
> between nodes (I was templating spreadsheets) and how node attributes
> were used.
>
> Maybe things have improved by now, but if I had to do the same thing
> again I would not rely on Libre Office. There's too much bloat to deal
> with on both the development and system administration sides.
>
> The big draw with Libre Office was that you can let business users
> easily customize the templates. A friend of mine is doing a SaaS web
> app in Common Lisp, generating documents via LaTeX, and is making a
> small web app that lets users customize the template layout. IMO this
> is simple enough to do that it's the approach I would use today (using
> CL-PDF or maybe cl-typesetting directly instead of going through
> LaTeX).
>
> Vladimir
>