lkml.org 
[lkml]   [1996]   [Apr]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectKernel Database [take 2]
Folks -

Yesterday I posted my announcement (and patch) for my Kernel
Database. Over the next 10 hours or so, I received somewhere between
100-120 hits on my page, but only two entries into the database.

I realize that the page isn't pretty and doesn't give a lot of
analysis, but that's simply because I haven't had the time to work on
the output.

The database itself is keeping track of quite a few things. It
measures:

- The kernels that are being used to compile others
- The versions of GCC that are being used and on what kenels
- How frequently people upgrade their kernels (taken from the
email address and what kernels they've registered)
- The low/average/high "jumps" people make between kernels
- The architecture of the machines running Linux
- The number of (unique) people participating

And some information on certain drivers/options:

- SCSI vs. IDE vs. both
- Modularization vs. compiled-in

Now some of this information is taken at "face-value". For example, to
determine the size of a "jump" between kernel versions, the program
takes the kernel version from "uname" and/or the last registered
kernel from the same email address.

As I said, unfortunately the only side of this that you are seeing
right now is the web page, and it's easily the worst one. I hope to
have all that corrected this weekend.

The patch itself I thought was fairly non-intrusive, touching only the
Makefile, scripts/Configure (to add a notice) and add the file
Documentation/kerneldb.txt.

When I wrote the Makefile patch, I tried to used a few tools as
possible (being scared of different permutations of people's
systems). I plan, however, to change that a bit and provide a little
more information. Luckily, when I wrote the original, I used field
delimiters to mark different areas; so now I can expand that
easily. One of the areas I plan to add is a marker to distinguish
between "test" kernels and "standard use" ones. That way we could also
keep track of whats being tested (and by whom) and how often.

I know that Linus (hi!) has (once again) put a feature freeze on
the kernel, but I was hoping to sneak this in anyway. I thought that
including this in the (early) production kernels would provide the
largest audience of people and therefore the most accurate
database. Additionally, it's not a feature :)

So there you go. I hope that everyone will participate in this, it's
the only way we can build an accurate database. Also, please don't
take the simple output as measure of the entire project, I like to
feel it's more advanced than that.

As always, questions/comments are welcome.

N!
__________________________________________________________________________
Nicholas J. Leon (pgp/finger) nicholas@binary9.net
Shinanyaku on Kali:Central descent2/warcraft2
Technical Development, Wachovia Bank real job!




\
 
 \ /
  Last update: 2005-03-22 13:36    [W:0.023 / U:0.420 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site