Re: Ajax based client for TDDB....
Dhruv Matani
Let me rephrase it:
List 3 benefits of using Ajax over normal HTML request/response [10M] :P
[1] The main being that the clients may not have anything loaded
except for a browser which supports Ajax(javascript, etc....). For
that matter, someone could just boot with a linux live cd, and connect
to the server using the browser.
[2] Lesser load on the server(compared to traditional HTML approach)
since formatting, etc.... is done at the client side.
[3] Looks good :P
If someone is prepared to take this up, I would be most interested in
it. Maybe I could even install it at college once it's done, because
the comp. dept. @ our col. lacks a DBMS lab. We used to visit the IT
dept. for that since, they are the only ones to run an _oracle_ server
machine. yes, an isolated box in a dark corner of the room ;-)
Neepa M'aam I understand that the juniors have to choose a project for
well, their final year submission purposes. If an when you make the
announcement(after the exams), please could you mention this as
something that could be worked on. Here are some details:
Someone who is interested in WT, and has a taste for designing
interfaces(Pascal).
Should know Javascript, basic SQL and wouldn't mind learning Ajax.
Regards,
-Dhruv.
On 6/3/06, Dhruv Matani <dhru...@gmail.com> wrote:
> Any takers????
>
> Here's a stray link I found:
> http://www.devarticles.com/c/a/XML/Building-an-AJAXBased-Chat-Interacting-With-a-Database/
> http://blogs.zdnet.com/BTL/index.php?p=1201
>
>
>
>
>
>
> --
> -Dhruv Matani.
> http://www.geocities.com/dhruvbird/
>
> "The biggest room is the room for improvement."
> -- Navjot Singh Siddhu.
>
--
-Dhruv Matani.
http://www.geocities.com/dhruvbird/
"The biggest room is the room for improvement."
-- Navjot Singh Siddhu.
Sandesh Singh
juniors, for things like security, user support etc.
Let's make a list of all such projects, and become a `mentor'.
(I plan to implement nested queries in vacations ... so keep that OUT :P )
--
sss
Dhruv Matani
>
> Hmm ... I was thinking of this too ... handing out sub-projects to
> juniors, for things like security, user support etc.
>
>
> Let's make a list of all such projects, and become a `mentor'.
hehehe:D lolz!!!!
It actually does make sense.
>
> (I plan to implement nested queries in vacations ... so keep that OUT :P )
gr8. I hope TCS doesn't mar our spirits :P
Actually, nested queries is a bit too much for a 2 month vacation,
because naming will be a real mess.
I propose we go ahead with:
[1] Non-field projections: SELECT x+y FROM tab1;
[2] Aggregates: SELECT MAX(x) from tab1;
SELECT MAX(x+y) FROM tab1;
[3] GROUP BY: SELECT MAX(x), y FROM tab1 GROUP BY(y);
[4] Other(better) JOIN algorithms. What I want to do is make a
cost+greedy JOIN selection algorithm, which doesn't eat up too much
time trying to find the _best_ join method. Currently, there are no
indexes in TDDB, so the JOIN process will have to do some
preprocessing like hashing or ordering of the input, which would
require that the full table be read in from disk.
[5] Another thing that come to mind is pushing down selections with
CONSTANT terms to select_helper_single_table.
[6] Make a plugin type of architecture for select_helper_single_table,
so that we can make TDDB read data from external data sources.
[7] Optimizations in the parser so that you can do redundant
expression elimination:
SELECT * FROM tab1 WHERE 1=1 AND 0=1; /* Always FALSE, because 1
branch of AND is const FALSE. */
SELECT * from tab1 WHERE 1=1 OR x=y; /* Always TRUE, because 1 branch
of OR is const TRUE. */
SELECT * from tab1, tab2 WHERE x=y AND x=23;
/* Push the projection down to selection of tuples from the disk. If x
is a unique field, stop when the first tuple is found. Also, rewrite
the query as:
SELECT * from tab1 WHERE x=23 AND y=23;
This form _should_ work for all queries of this form. */
-Dhruv.
>
> --
> sss
Pascal Nunes
1. Fix existing flaws... there are a hell lot of them
2. Automate the process of adding new nodes.
3. put a replicator over the distributer
Dhruv Matani
> Think that I have to do...
>
> 1. Fix existing flaws... there are a hell lot of them
>
> 2. Automate the process of adding new nodes.
>
> 3. put a replicator over the distributer
Ok, great, so we all know what needs to be done.
Btw, how are u placed for the summer? I mean any plans? Cause we are
planning a trip to Pune. I mean the original group that went last
time. Maybe some more. Are u in?
-Dhruv.
Pascal Nunes
On 6/3/06, Pascal Nunes <pascal...@gmail.com> wrote:
> Think that I have to do...
>
> 1. Fix existing flaws... there are a hell lot of them
>
> 2. Automate the process of adding new nodes.
>
> 3. put a replicator over the distributer
Ok, great, so we all know what needs to be done.
Btw, how are u placed for the summer? I mean any plans? Cause we are
planning a trip to Pune. I mean the original group that went last
time. Maybe some more. Are u in?
Sure! I have no other plans for this summer. Lets meet soon after the exams... so do a little planning, and a lot of celebrating the end of engineering.
-pascal.
--
Any tool is only as good as it is used.
.:[ http://www.pascalnet.tk ]:.