glue-wg-bounces@ogf.org [mailto:glue-wg-bounces@ogf.org] On Behalf Of David Horat said:
Finally I have put together all major design decisions made during the LDAP implementation process and created a document with them. It is now open to your comments and suggestions.
I think the most obvious thing is that the DIT diagram is wrong, and also the format is confusing - it isn't a tree! i.e. in an LDAP tree every node needs to be an object, and all the lines need to be basically vertical. Also you don't mention the grouping objects or say what constraints exist on the DN. And the example DN is completely wrong, there are no EntityIDs! Also more generally I think the explanations need to be made clearer, and to give examples - maybe you should try this on someone who hasn't been involved with glue 2 up to now, I doubt they will understand the current version. Also as a general point it seems that none of the names in the text have the GLUE2 prefix. You mention the foreign keys, but don't say what principles we used to decide which end gets the key, or explain how the keys can be used. Indeed you don't even say what the key actually is (i.e. the ID). You don't seem to mention case sensitivity, and don't explicitly say that optional/mandatory and single-multivalue sttributes are reflected directly in LDAP. Also we still have the pending question of whether we want to use DirectoryString (UTF-8) rather than IA5String. There is no reference to the actual LDAP schema. Stephen -- Scanned by iCritical.