Ideas for a project? (a MSc project, sort of)
Jason Cunliffe
jasonic at nomadicsltd.com
Thu May 3 18:17:32 EDT 2001
PROJECT PROPOSAL: 'PyKBASE' or "Stranding on the shoulders of very smart
giants"
Show them Kdb from Kx systems and ask them to try to implement something
like it in Python.
Almost all the parts probably exist .. how to put them together
intelligently and get it to fly..?
http://www.kx.com
http://www.kx.com/download/latest_kdb.htm
[more links at the end]
Phase 1. INSPIRATION
Initally, don't ask them to port the original, instead try to get them to
analyze and be insprired by it. How should/could/should it work in Python,
pros cons improvements differences etc..
However, if that job is too overwheleming then go swiftly to PLAN B = write
really nice Python bindingts and GUI for Kdb instead :-)
Phase 2. SPECIFICATION
ok so its a huge challenge... help them to focus towards what they think can
be achieved..
tiny
very fast [that cuold be hard, but after smart design optmiiztion would be
last phase]
easy to install
cross-platform
RDBMS sorts on Columns not Rows
TimeStamp indexing lookups [MSEQDICT + mxDateTime]
Embedded webserver [run KDB and then login from your browser = wow!]
Gui [wxPython = nice grids/tables/activex embedding]
Additional features to include not in Kdb:
Embedded nntp server/client
This would allow the actions to the database to be recorded as a news
thread. But also using a simple minimal tag XML protocol one could 'post'
data, changes. methods to the database.
Cool better stuff of their own !!
Phase 3. DESIGN + IMPLEMENTATION
Python allows one to thinkas one works right..
Phase 4. DOCUMENTATION
This shoudl be happenign all along butprobably wnot. Using Docstrings
,Keyargs adn other neat featues yout shuodl be able to design for the
project be reasonably self documenting
Phase 5. OPTIMIZATION
identifying where how - anaylisis, tests, understanding dependancy of
system..
factoring out loops, perhaps some C..
Phase 6. GROWTH
Note that Kdb speed depends in part upon doing come very simple things like
using memory.
Also Python's deep use of Dictionaries should be very helpful
PS> I am nothing affilaited with Kx systems - I just think what they are
doing is very cool, makes a great project, and would love to see it happen
in Python
SOME USFUL LINKS if this interests you:
But for starters three are several PDF documents here..
http://www.kx.com/developers_white_papers.htm
And one can always get usually get a quick heads-up for real-world use of a
technology by looking at its list archives:
http://www.kx.com/listbox/
[links to both K and Kdb lists]
some [slightly random] interesting thread topics..
http://www.kx.com/listbox/kdb/msg00395.html
http://www.kx.com/listbox/kdb/msg00418.html
http://www.kx.com/listbox/kdb/msg00509.html
http://www.kx.com/listbox/kdb/msg00672.html
http://www.kx.com/listbox/kdb/msg00707.html
http://www.kx.com/listbox/kdb/msg00787.html
http://www.kx.com/listbox/kdb/msg00815.html
http://www.kx.com/listbox/kdb/msg00968.html
http://www.kx.com/listbox/kdb/msg00986.html
http://www.kx.com/listbox/kdb/msg01055.html
http://www.kx.com/listbox/kdb/msg01091.html
http://www.kx.com/listbox/kdb/msg01094.html
http://www.kx.com/listbox/kdb/msg01253.html
http://www.kx.com/listbox/kdb/msg01266.html
http://www.kx.com/listbox/kdb/msg01318.html
http://www.kx.com/listbox/kdb/msg01320.html
http://www.kx.com/listbox/k/msg03376.html
http://www.kx.com/listbox/kdb/msg01382.html
http://www.kx.com/listbox/kdb/msg01384.html
http://www.kx.com/listbox/kdb/msg01425.html
concise documents in http://www.kx.com/a/kdb/document/
http://www.kx.com/a/kdb/document/acid.txt
http://www.kx.com/a/kdb/document/paradigm.txt
http://www.kx.com/a/kdb/document/io.txt
http://www.kx.com/a/kdb/document/contention.txt
http://www.kx.com/a/kdb/document/foreign.txt
http://www.kx.com/a/kdb/connect/http.txt
http://www.kx.com/a/kdb/connect/kdbc.txt
http://www.kx.com/listbox/kdb/msg01045.html
wall street consultants
http://www.mkrsw.com/
PRIMER
http://www.mkrsw.com/doc/primer.html
Time Series in Finance: the array database approach
http://www.cs.nyu.edu/shasha/papers/jagtalk.html
and for desert
http://citeseer.nj.nec.com/408217.html
http://www.maverick-dbms.org/
hth!
--Jason
___________________________________________________________
Jason CUNLIFFE = NOMADICS['Interactive Art and Technology']
----- Original Message -----
From: <Richard at volpato.net>
To: <jasonic at nomadicsltd.com>
Sent: Thursday, May 03, 2001 2:55 AM
Subject: Query about KDB
> Jason,
>
> Quick query, just followed up your reference to KDB. Interesting stuff,
particularly regarding its use of time as part of records. Can you point me
to papers on KDB deployments or comparisons with other databases?
>
> Thanks
----- Original Message -----
From: "eloy" <eloy at rabinf50.uco.es>
Newsgroups: comp.lang.python
Sent: Wednesday, May 02, 2001 4:20 AM
Subject: Ideas for a project? (a MSc project, sort of)
So, as I am new to the Python community, I would like to ask
if there is any idea about projects that could fit that description. Of
course the results would be available to everybody.
More information about the Python-list
mailing list