X-Git-Url: http://lists.indexdata.com/cgi-bin?a=blobdiff_plain;f=TODO;h=0e05d0ccfb575a53b6583bef6e46b495f66cf1b1;hb=7ce18cf5c766487c184322478d9a656ac63e1456;hp=09081de0e68f1e330b05befaec2cc25c91844287;hpb=819a1eeaa1f324f61d5549768cc51ab7a45068f1;p=pazpar2-moved-to-github.git

diff --git a/TODO b/TODO
index 09081de..0e05d0c 100644
--- a/TODO
+++ b/TODO
@@ -1,25 +1,18 @@
 Short term: 
 
-Structured target profiles -- Zeerex.
-Normalization of records -- test if libXSLT is fast enough to:
-  -- Extract Subject headings, titles, dates, authors for facets and display
-  -- Merge keys for relevance ranking, etc.
-  -- Also normalize to UTF-8.
-Implement timeout of sessions, so they don't sit there forever (simple).
+Pull more metadata from records. Figure out how this relates to record retrieval
+Structured target profiles -- Zeerex?  -- requires libxml
 Sort by title, date, author. Parameter to 'show' webservice command.
-Additional facets. At least do 'author'.. think about making it general.
-Support for 'blocking' requests, at least for record display before first records
-   are retrieved (to reduce polling).
-Implement detection of 'search complete', so client knows when to stop polling.
-Full record retrieval.
-Implement hitsbytarget function to emulate traditional LOT-style U/I
+Full record retrieval. -- libxml
 
 Longer term:
 
-Server-specific CCL-mapping (simple)
+Server-specific CCL-mapping (simple, follows from structured target profiles)
 Networked mechanism to control configuration
 Target selection by metadata -- 'subject', 'location', 'library
 	type' -- should be totally generic to support requirements of
 	different projects.
+Factor Z39.50 stuff out from pazpar2.c to separate file to make room for
+   later SRU implementation (essential if we do THDI or LOT. Not hard to do)
 
 Nice to have: