Re: "image analysis heuristics" (ISSUE-66)

On 2/7/10 10:19 AM, Julian Reschke wrote:
> On the other hand, there's also lazy spec writing (in general, just not
> over here). For instance, many authors - when they link to other specs -
> apparently are too lazy to make their references precise (why should I
> add a section reference when the reader can figure that out him/herself?)

Worse yet, in the case of web-related specs, much of the time the 
linked-to spec is not compatible with the spec that linked to it. 
Sometimes as a matter of pragmatics, sometimes as a simple logical 
inconsistency due to using the "main idea" of the linked-to spec but 
changing all the details (URIs in HTML4 are an example of the former; 
XLink in SVG an example of the latter).  These problems is so widespread 
in web-related specs that the return on time invested in reading the 
linked spec is, in my experience, negative on average.  That is, reading 
it actually makes it _harder_ to implement the linking spec "correctly".

-Boris

Received on Sunday, 7 February 2010 15:52:19 UTC