Hi Alex,
For most topics or areas to document or describe, the most
important thing to get right is the organization of the data. I've
seen a number of wikis that contained huge quantities of
information - but I could never find what I wanted without a lot of
effort. Wikis make it easy to add new categories - so it's easy to
have too many categories, or individual articles mis-categorized.
That's an excellent point... As the ontology manager proves, I am
trying to find a way to back up my own wiki with an ontology. In a
way, the idea is to solve the salesman problem. The user should be
provided with many different paths to the same information and should
always be able to accede the information that is relevant to him in
less than 5 clicks. Slight differences in word use (for instance
Transcript vs xTalk) should be irrelevant.
Ultimately, what I am after is to be able to use a single
classification system to acceded all bookmarks I have on the web, all
files I have downloaded from the web, as well as all documents I have
on my desktop computer.. transform my desktop into a (cleverly
organized) CMS... or rather treat my desktop and the web as part of
the same work environment (I have Broadband at home). When I would
look for a file on a given topic, I would be told of the related
topics higher and lower in the tree. I would be told as well of other
resources (images, rev projects, bookmarks, Contact Addresses,
Emails, etc.) that belong to the same category.
Duplicating spotlight??? Not really... spotlight suffers exactly from
the problem highlighted by Alex. It let you randomly search for
documents, based on their content. For sure, you can without too much
difficulty attach ad-hoc spotlight comments to anyfile (for this, you
can define in automator a finder action on a folder). But still, for
the research to be efficient, what is needed, is a kind of Trove
classification system to specify these keywords (http://
sourceforge.net/softwaremap/trove_list.php). With the way it works,
if you happen to change any keyword in the future (from Transcript to
xTalk)... your search fails.
A demo of my solution so far:
go url "http://revolution.lexicall.org/paypal/Download%20Manager.rev";
(sorry, this one is password protected, still working on it)
The major features are:
(1) URL+Page title are added by drag and drop of the URL from safari.
(2) Keywords are easy to add, thanks to a system which allows for the
autocompletion of keywords (so to guide the use in the provision of
correct keywords). Alternatively, you can look for the keyword in the
keyword list and double click on a line of the list to have it added
to the entry's keywords.
#######################################################
##
## Automatic Keyword completion routine
##
## Constraints:
## Keywords are entities separated by ";". Eventually, they
can be separated by "; ".
## A start of line and end of line also marks a start or end
of a keyword. Spaces are
## are allowed within a keyword
## Behaviours:
## 1. A key is pressed in the keywords field
## If this key is not a space or a keyword separator,
then completion is proposed.
## 1) information about the start and end position (in
characters) of the keyword
## is retrieved.
## 2) a list of the keywords that match the start of
the letters already provided
## is generated (for this, the full list of keywords
is used, field "lst_keywords").
## 3) This list is sorted by alphabetical order
## 4) By default, the first item in the list is
proposed as completing segment
## 2. The up or down arrow keys are pressed
## 1) The next item in the list of completing segments
is used to complete the letters
## appearing between the start of the keyword and
the selection point
## 3. The left or right arrow keys are pressed when the
shift key is down
## 1) Jumps to the start (left arrow) or end (right
arrow) of the keyword
(3) keywords are easy to find even when you only have a partial match:
If you do a mousedown on the arrow part of the combobox under the
list of keywords, it populates the Combo-box with the list of
keywords in lst that match the selectedtext in the combobox.
Next step is to migrate it to SLQlite, to allow access from the web
as well as from my desktop... The thing to know is that you can
easily manipulate SQLlite from the shell. SQLlite itself can be
downloaded for free for many platforms (http://www.sqlite.org/).
Then (sorry for this, Chipp, but an easy and free db management
system may be decise for bringing new users to revolution), it
suffices to write some functions like this:
########################
##
## SQLlite
##
on updateLstFromDB
put getPanelData() into tA
set the defaultfolder to "/users/mlange/Work/Scripting/
___myProjects/sqlite/"
put "select * from downloads;" into tSQL
put shell("./sqlite downloads.db '" & tSQL & "'") into tData
replace "|" with tab in tData
put tData into field "lst_downloads"
end updateLstFromDB
on deleteFromDB tID
put getPanelData() into tA
set the defaultfolder to "/users/mlange/Work/Scripting/
___myProjects/sqlite/"
put "delete from downloads where id=" & tID & ";" into tSQL
put shell("./sqlite downloads.db '" & tSQL & "'") into tData
-- updateLstFromDB
end deleteFromDB
on addEntryinDb
put getPanelData() into tA
put the internet date into tDate; put tDate; convert tDate to
dateItems;
put item 1 of tDate & "-" & item 2 of tDate & "-" & item 3 of
tDate && item 4 of tDate & ":" & item 5 of tDate & ":" & item 6 of
tDate into tTimeStamp
put quote & tTimeStamp & quote, quote & tA["keywords"] & quote,
quote & tA["title"] & quote, quote & tA["url"] & quote, quote & tA
["desc"] & quote into tData
set the defaultfolder to "/users/mlange/Work/Scripting/
___myProjects/sqlite/"
put "insert into downloads (ID, date, keywords, title, url,
description) values (null," && tData & ");" into tSQL
replace numtochar(13) with " " in tSQL --- Could escape them,
but I prefer to avoid cr in my application.
replace cr with " " in tSQL
replace " " with " " in tSQL
put shell("./sqlite downloads.db '" & tSQL & "'")
-- updateLstFromDB
end addEntryinDb
... At least, that's fine for personal use (that's early code, it
works, but not really seriously tested). For professional projects,
to guarantee perfect cross-platform compatibility, best is to use
Chipp's solution (altSQLlite - http://www.altuit.com/webs/altuit2/
altSQLiteCover/default.htm).
Are other persons on this list interested in this kind of problem/
solution?
Marielle
------------------------------------------------------------------------
--------
Marielle Lange (PhD), Psycholinguist
Alternative emails: [EMAIL PROTECTED], [EMAIL PROTECTED]
Homepage: http://homepages.lexicall.org/mlange/
Lexicall: http://lexicall.org
Revolution-education: http://revolution.lexicall.org
_______________________________________________
use-revolution mailing list
use-revolution@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-revolution
