[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: local editing, heretical views



On Wed, 4 Mar 1998, Erik Ostrom wrote:

> of course, opaque identifiers aren't always even what you _want_.
> in my usual hacking client, i can take a description or a piece of
> code i'm editing, and easily retarget at a different object, update it
> to reflect a verb name change, or even send it to a different mud.
> (this is especially handy when i have a wizardly and a non-wizardly
> character on the same system; i often get an editing buffer via my
> non-wizardly connection, then realize i need to use the wizard to edit
> the text in question.)  it's difficult to do any of these if you hide
> from the user the state that says where a given piece of text is
> going.

Quite..
> 
> from a software engineering standpoint these are gross things to be
> able to do, but for practical use, i'd be sad to see them all go away.
> and yes, you can always cut and paste.  but with round trips involved
> to get a new editing buffer, it can be significantly slower to do so.

Aggreed..
> 
> anyway, i guess i'd prefer that "ident" which must be identical from
> send to reply by replaced by "location" that is just another field
> that the user can munge.  for the simple protocol, i mean.  for a
> complex protocol with persistent identifiers, we can provide other
> ways of changing what a session points to.

I would suggest requiring ident OR location, you don't need both but you
need one or the other..
If someone designs a system based on the spec which does not allow
changing of the location then fine, but this allows simple systems based
soley on location.. 

Zephaniah E, Hull
> 
> --erik
>