!!!Risten2 Meeting 4.6.

Attending:
* Børre
* Tomi
* Sjur

Topics:
* incremental search
* demo applications & prototypes
* next meeting with targets

!!!Incremental search

Tomi found the old code for the previous incremental search prototype, which is based on [script.aculo.us|http://script.aculo.us]. It seems pretty easy to use and integrate with XQuery.

eXist is using the [ACE javascript framework|http://ace.ajax.org/api/ace.html], and [jquery|http://jquery.com/].

Incremental search requirements:
* be able to delay search until a certain amount of chars are types
* set delay time before searching (ie wait 0.5 seconds before searching)
* display options:
** in-line ("autocomplete" in the search box)
** hit list below (= list of matching index entries in eXist) - this should always be visible to guide the users, so that they can see what the possible lookup words are
** possibly a combination of both
* spell-relax/fuzzy search option?

!!!Demo applications & prototypes

Data organisation:
{{{
data/
     dicts/ <- this in webapp svn
           Kintel-SMJNOB/ <- but not these ones (except for one development set)
           ...
     lists/ <- this in webapp svn
           fellesliste/ <- but not these ones (except for one development set)
           ...
     terms/ <- this in webapp svn
           mekanikk99/ <- but not these ones (except for one development set)
           ...
}}}

Data (dictionary/terminology) dirs could be soft-linked into the data dir of the Risten2 data dir when ready, one directory at a time. But first we need to define the DTD/schema of the Risten2 dictionaries/terminologies, and see how it would/could fit the existing schemas of the dictionaries.

About {{lists/}}: these are very simple bilingual, 1:1 lists for specific domains, such as world geography for SMA elementary school, where the requirements for dictionaries and terminology collections place too high an overhead on the data in relation to the target user group and the data amount at hand (typically less than 100 entries). The important aspect is to get data out fast and searchable. We will experiment with lists for the prototypes, and see how it all works out.

The Risten2 web application should be stored in svn in:

{{{
$GTHOME/apps/risten2
}}}

!!!Next meeting

* Next Monday (11.6.) after lunch

Targets for next meeting:
* incremental search integrated with one of the prototypes
* svn development cycle established
* development data:
** dicts: smanob
** lists: fellesliste
** terms: mekanikk99