httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <>
Subject Re: Log Ignore Hosts patch uploaded
Date Mon, 26 Feb 1996 23:37:08 GMT
On Sun, 25 Feb 1996, Adam Sussman wrote:
> > Under normal circumstances, that is an extremely bad idea.  Is there a 
> Why is this a bad idea?  I admit that this just a 'feature' and that
> there seems to be some amount of resistance to feature enhancements here,
> but I don't see what's wrong with it.  Its an option, not an irrevocable
> behavioral change.  You can choose not to use it and the code isn't going
> to hurt anything by simply being there.

One of the key elements in good software engineering is knowing how to 
avoid "feature creep".  One of the best ways to avoid it is to not 
implement certain features, no matter how easy it would be, which would 
engender in managers or users the expectation that more features "like 
it" would be implemented.  If a LogIgnoreHosts directive was put into 
Apache, I do not think it would be a stretch to hear calls fo a 
LogIgnoreObjects directive, a LogIgnoreResponseCodes, a LogIgnoreImages, 
etc etc, since the conceptual step from LogIgnoreHosts to any of those is 
much smaller than the conceptual step from LogFormat to LogIgnoreHosts.  
We could collectively decide that, sure, it's right and appropriate to 
put a whole bunch of logfile processing intelligence directly into 
Apache.  But I think that would contradict with other goals, such as 
keeping things simple and efficient.  Fortunately there is already an 
extremely good tool for logfile analysis, Perl, and to write the 
equivalent of "LogIgnoreHosts" for a perl engine to process the logfiles 
is about 2 lines of perl code.  

I think the general consensus has been to leave logfile manipulation to other
programs and processes - that's certainly what I'd prefer.  The LogFormat
directive is about as far as it needs to go, and even that can include a pipe
to another Unix process.  Now, proxies will need some extra intelligence if
it's going to support any report-back mechanisms the W3C elects to pursue,
but that's a separate issue. 


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--  http://www.[hyperreal,organic].com/

View raw message