1. Internet Publication of Geneva Justice Decisions A case study laurent.dami@justice.ge.ch

2. LD, PJ-GE, july 20062 Agenda u context presentation u justice.ge.ch/jurisprudence : short tour u technical information u some lessons about Perl in the enterprise

3. Context presentation

4. LD, PJ-GE, july 20064 A justice decision u is a structured document header / facts / law / conclusion may have a 2 nd, anonymous version u has a unique identifier (e.g. ACJC/123/2005) u has a context (metadata) date / names / topic / keywords / summary / etc. u is archived into a collection minutes du TA / CJC / TPI / etc.

5. LD, PJ-GE, july 20065 Lifecycle receive case investigate write project deliberate finalize send supply context archive clerkjudgecollege

6. LD, PJ-GE, july 20066 Electronic archive : requirements u store document multiple formats fulltext indexing u store metadata structured fields quick search (unstructured!) u intelligent presentation automatic hyperlinks u offline / CDROM copies per collection

7. LD, PJ-GE, july 20067 Some figures u Intranet : 20 – 30 collections Internet : only 2 collections for the moment u 500 to 50000 decisions per collection for about 10 years of data u 2 – 50 pages per document

8. LD, PJ-GE, july 20068

9. 9

10. Short tour http://justice.geneve.ch/jurisprudence

11. LD, PJ-GE, july 200611

12. LD, PJ-GE, july 200612

13. LD, PJ-GE, july 200613 metadata search fulltext search

14. LD, PJ-GE, july 200614

15. LD, PJ-GE, july 200615 Qualité pour agir

16. LD, PJ-GE, july 200616

17. Technical information

18. LD, PJ-GE, july 200618 Which kind of solution ? u Electronic Doc. Management System not well suited for multiple disjoint collections approval / workflow not relevant u Database many fields : too much structure for easy searches (SQL not well suited)  see CPAN SQL::KeywordSearch !

19. LD, PJ-GE, july 200619 Storage of a collection metadata.txt words.bdb file.{doc,html,pdf} w2docs.bdb positions.bdb fulltext index in BerkeleyDB format documents flat file config.txt

20. LD, PJ-GE, july 200620 Phases for a search Parse request Metadata searchFulltext search Merge results Sort & slice Contextual excerpts Display

21. LD, PJ-GE, july 200621 Main Modules Search::QueryParser Search::Indexer BerkeleyDB Template toolkit File::Tabular::Web File::Tabular CGI AppConfig not (yet) on CPAN ModPerl::Registry

22. Some lessons about Perl in the Enterprise

23. LD, PJ-GE, july 200623 Context : Geneva Justice u finished phase 1 (collaborative software, document management) u ongoing phase 2 : rewrite the old COBOL application for case management using mod_perl Catalyst DHTML + Ajax u smooth transition COBOL and Perl must live side-by-side for several years

24. LD, PJ-GE, july 200624 Acceptance u strong internal resistance bad image : low-tech, hacking, scripting Perl5 features not known  objects, namespaces, closures, etc. not "standard" (i.e. not Java) fear not to be able to maintain and industrialize cheap means "not serious" u but: Perl productivity wins !

25. LD, PJ-GE, july 200625 Perl Job Market u found more people than expected. But all coming from US / UK used Perl several years ago, now on Java / PHP missing other skills (modeling, communication, project management) u apparently not enough "average profiles" few top stars many low-level geeks Perl not taught at school !

26. LD, PJ-GE, july 200626 Industrialization u Release management : granularity mismatch production guys want  big tarballs  few updates  strict release process development guys want  small and frequent updates using cpan / cpanplus / minicpan  fast release process, short feedback loop

27. LD, PJ-GE, july 200627 Development u TMTOWTDI yet developers need guidance  many thanks to Damian Conway !  IDE CPAN : how to manage proliferation  cpanratings not exhaustive/reliable enough  which modules were rated by ?  inverse dependencies