This just immediately made me think "Oh, wouldn't it be amazing if someone did something similar with python-ldap, so you could navigate the DIT and farm entries via a TiddlyWiki-based UI..." Just about all the generic LDAP Manager tools suck, and the app-specific management tools tend to be even worse, e.g. the things hacked together to craft/poke mail or UNIX account related entries. You tend to end up faced with either a grim UI that doesn't quite manage all the attributes you need, else hand editing an LDIF file in vi.
But, yeah, nice demonstration of the power of having a pluggable store architecture. I'd previously only considered file vs. RDBMS etc for plain old data storage.
Jay Fienberg
· 9 months ago
The web over IMAP, a subject near and dear to my brain!*
For caching, I wasn't clear from the description whether the performance hit was from downloading IMAP messages over the network, or just from parsing locally stored IMAP boxes? If it's just the former, then maybe you'd want to think of the IMAP boxes the way a mail client would, e.g., building up a local copy that can be sync'd with the server.
* I wrote a rambling thing about about the web over email back in the day - still interested in the implications of "from," "to" and "cc" with regards to a linkable web of information--maybe some food for thought:
It's the former. TiddlyWeb needs to get at the same information many times over in any typical session, and in the existing code there is no local synching copy: there's just the view of what's on the IMAP server. So if we want to look at a mailbox again, we look at the whole mailbox again.
Local boxes would be useful. I've also got some work started on a generic cachingstore which wraps any other TiddlyWeb store with a cache. For remote resource handling cache validations gets messy, but there's always a way!
Andrew Back
· 9 months ago
Right, so the mail client Mutt is slow loading my Maildir without caching. A generic cachingstore sounds like the thing, but would it be configurable to cache only headers/tiddler names and possibly tags?
I'm thinking I really do need to learn Python. First GNU Radio, and now I'd really love to be able to implement a TiddlyWeb ldap store...
BillSeitz
· 9 months ago
Do you define a sister-wiki-space, so you can link wikiwords in emails to wiki pages?
cdent
· 9 months ago
Not yet, but this is definitely something I have in mind. The code above is very much just raw ick, not a lot of pretty going on, but some kind of sister linking would be very valuable.
Similarly my work with multiple recipes in the same TiddlyWeb setup, but different TiddlyWikis is making it clear that there needs to be syntax for a kind of intra-TiddlyWeb interwiki links.
All Comments
This just immediately made me think "Oh, wouldn't it be amazing if someone did something similar with python-ldap, so you could navigate the DIT and farm entries via a TiddlyWiki-based UI..." Just about all the generic LDAP Manager tools suck, and the app-specific management tools tend to be even worse, e.g. the things hacked together to craft/poke mail or UNIX account related entries. You tend to end up faced with either a grim UI that doesn't quite manage all the attributes you need, else hand editing an LDIF file in vi.
But, yeah, nice demonstration of the power of having a pluggable store architecture. I'd previously only considered file vs. RDBMS etc for plain old data storage.
For caching, I wasn't clear from the description whether the performance hit was from downloading IMAP messages over the network, or just from parsing locally stored IMAP boxes? If it's just the former, then maybe you'd want to think of the IMAP boxes the way a mail client would, e.g., building up a local copy that can be sync'd with the server.
* I wrote a rambling thing about about the web over email back in the day - still interested in the implications of "from," "to" and "cc" with regards to a linkable web of information--maybe some food for thought:
http://icite.net/blog/200605/movable_web.html
Local boxes would be useful. I've also got some work started on a generic cachingstore which wraps any other TiddlyWeb store with a cache. For remote resource handling cache validations gets messy, but there's always a way!
I'm thinking I really do need to learn Python. First GNU Radio, and now I'd really love to be able to implement a TiddlyWeb ldap store...
Similarly my work with multiple recipes in the same TiddlyWeb setup, but different TiddlyWikis is making it clear that there needs to be syntax for a kind of intra-TiddlyWeb interwiki links.