Thanks all.
Yes, MD5 is unreliable. But it has proven effective in identifying
1,000s of duplicate profile images in spam accounts before a single
byte change throws the checksums off. Looking at the path is certainly
an option, but I was worried that these may change suddenly without
warning.
This is a pretty hostile worded email for someone who is asking for help for
a problem that isn't necessarily directly related to the API.
Just saying...
Tim,
Twitter deploys dozens of code branches each week, most of which
probably contains at least a few user visible changes. The changelog
is difficult enough to follow internally. Externally, it would be
hopeless. Notifying on each and every change isn't a tractable
problem.
Although there is
Factoid, FWIW: so far, I've found 7:
http://s.twimg.com/a/1252980779/images/default_profile_x_normal.png
where 0=x=6.
Jim Renkel
-Original Message-
From: twitter-development-talk@googlegroups.com
[mailto:twitter-development-t...@googlegroups.com] On Behalf Of
timwhitlock
Sent:
I don't think it sounded hostile, and it sounded to me like he was proposing
it be part of the API, which I agree. That would be pretty useful
information, especially in a constantly changing environment.
Jesse
On Tue, Sep 15, 2009 at 9:52 AM, Adam Cloud cloudy...@gmail.com wrote:
This is a
Tim,
We specify full URLs to images so that developers don't have to supply
custom code to pull in profile images and background images. It sounds
like you have a pretty unusual use case for our profile images.
For what it's worth, I think we deployed six variations of those
images, but our
I have not looked at this so this is mostly curiosity.
Why use md5 on a moving target? Who knows when someone may resave an
image to compress it more.
I bet 1% compression savings translates to thousands of dollars over
short time.
Isn't the path relatively static?
/images/default...