DASL ID 3 Apr: Re Ch 4 searcherror

I suggest DASL not attempt to define searcherror tags (chapter 4), because
I think the effort to standardize and implement this outweighs the value.

Effort: It will take time to define a useful set of error codes, to
document them, show how they map to various cases, and so on.

Value: It will not buy any additional interoperation.  The most it can
possibly do is allow a UI using DASL to display a better error message than
one that would be included in the reply with the 400.  I admit, there is
some value in this, particularly for non-English speakers.  But there is
not enough value considering the cost to DASL to define and implement it.

In fact there is some anti-value as well: Mandating that the 400 reply
ALWAYS be an XML document means that every compliant implementation must
use it.  This will prohibit use of HTML (human readable) error replies. 

We have plenty of hard work to do just to define a core set of operators, a
grammar, the discovery mechanism and making it all international.  Can we
drop this please?

Received on Thursday, 16 April 1998 13:42:58 UTC