Plans for
the future
We'd be particularly grateful
for contributions in the fields below:
Structured documents
A new XML parser/validator is in order. Important features:
- Full TT/MT compatibility.
Use markup trees internally. This is the primary reason for making a
new XML parser, since it allows for the use of token tree features like
the database (DB), network exchange and caching, streaming, etc.
- Subtree validation.
Enables you to validate document fragments by themselves. Very attractive
in an interactive environment.
- Reverse validation.
To provide lists of legal elements in a specific context. Also very
attractive for interactive applications, as it will enable you to present
lists of valid choices, etc.
Flux already comes with the RXP
parser integrated, and this can be further modified to suit our needs (use
token trees instead of the current internal structs).
Protocols/storage formats
- Unicode. A standalone
Unicode string handling layer, duplicating the <string.h> and
<stdio.h> 8-bit functions. Could be named strcpy16(), strlen16(),
etc.
- Distributed TTs.
Token trees shared between threads running on the same machine, as well
as remote ones. Need a special network layer for managing locking, transmission
and general bookkeeping. Good for multiprocessing, multi-user collaboration,
redundant clustering and whatnot. Some parts are in place already -
notably external TT data.
- Newsfeeds. NNTP.
Enough said for now.
- IMAP. Maybe POP3
too.
- IRC. Already implemented
as a blocking layer on top of SOCKs, but not checked in yet, due to
neccessary adaption/cleanup. Should queue and return token trees as
communication blocks, containing enough information that they can be
shuffled on through a COMM directly. Needs its own PROXY plug. Kinda
low priority.
|