Im writing a web tracker/logger

help required

[drewst]

Hey guys, i have many years experience in programming and am thinking of making a useful tool for my new found hobby of website design.

I want to develop a freeware website tracker/logger and stats application.

I have a few design queries though first and maybe some of the web masters on here could help me :)

1) I host my sites on a web hosts server not mine own YET :) so i have to consider that i dont have full admin rights.

2)Shall i write a script that is called on each visit to the site? if so what shall i write it in? javascript? php? .net? asp? my initial thoughts are PHP purely for compatibility issues (some web hosts still dont support asp, some borwsers ban java etc..)

3) What kind of info shall i log to be useful
1) Hits
2) IP address of visiotr
3) search engine keywords used
4) robot filertering for accurate stats results
etc... please suggest

4) How shall i record this info? initial thoughts are mysql simply for the power of queries hoever flat files may have an advantage i cant see :)

5) using databases allows the front end to be completely seperate from the storage of info and if i also allow people to enter sql they can have the power to query info that i may not be displaying.

HELP i need as many suggestions as possible, if anybody has and code snippets / advice or tip it wouls be much appreciated and of course a mention of you in the Source and developers documentation

OH one more thing
can anybody think of any snazzy names to call the application :)

Thanks

Drew

[Seige]

DrewStats - It's all about visuals.

How's that? ;)

[cgrantski]

These are simple looking questions with endless answers. If you expect others to use and like your freeware, in order to make the right decisions, I suggest that you read through this forum, read a couple of books on web analytics, and also start using software that's out there now - use it, not just look at it. Otherwise you might just be reinventing something that's already out there as freeware, or you might be making the same mistakes somebody else already made and not really using your energy to improve things.

[larryhatch]

Here's a separate but related idea. Some might like to sift thru their
access log files manually, but are put off by all the junk shown.

How about a utility that STRIPS OUT unwanted fields?

In my case, I just want to see date, time, what files were called up,
by whom or what, name of search engine if any (abbreviated GG=Google, YH = Yahoo ..) any search phrase, etc.
What I DON'T need is ie=utf-8, Opera 7.60, and a jillion other browser
versions and options. I don't care if the browser allows dirty pictures.
I hate those endless strings of garbage some records leave.

Here is what is left after I edited part of a log:

66.249.71.xx 23APR05 02:26:29 /MSTABDNA 2772 GGbot/2.1
64.136.49.xx 23APR05 02:26:31 /THEMEMAPS 7458 LH/SATL
64.136.49.xx 23APR05 02:26:32 /EMPTYQSL.gif 7515 LH/THEMEMAPS
66.249.71.xx 23APR05 02:26:43 /LTRSAM 304 GGbot/2.1
64.136.49.xx 23APR05 02:26:50 /LOSTLOCS 16589 LH/THEMEMAPS

I appreviated my base URL to just LH, and shortened the dates.
I deleted .html entirely, removed all "http:// and quote marks etc.

Fields to strip should be left to the user, and most remaining fields
can be shortened. I find that a lot easier on the eyes. -Larry