Re: gnats/437

[Yngve Svendsen]

At 13:56 25.11.2002 -0800, Zack Weinberg wrote:

> address@hidden writes:
>
> > Thanks for the report. GNATS 4 is a complete rewrite and no
> > compatibility between GNATS 4 tools and GNATS 3 servers and vice
> > versa can be assumed. This should be mentioned explicitly in the
> > documentation. I am setting this to analyzed now, and will update
> > documentation as soon as I can.
>
> That is unfortunate; it means I have to have two different versions of
> the tools installed, if I need to use both GNATS 4 and GNATS 3
> servers.  Also, the GNATS 4 tools have considerable usability
> improvements when it comes to accessing remote servers (3.x tools are
> locked to a single server at compile time, as far as I can tell).
>
> I don't see any reason in principle why the newer tools couldn't
> implement the older protocol for backward compatibility.

I think there are two main reasons why this hasn't been considered:

- Development resources: there are very few people actively working on 
GNATS development. The focus is squarely on getting a usable version 4 out 
of the door.
- A clean break with past: The GNATS 3 network protocol is badly designed 
and contains field-specific commands. query-pr for instance still supports 
the old field-specific options, but these are marked as deprecated and in 
the long run we will probably remove them in favour of the far more 
powerful query expression syntax. This will reduce code complexity a lot.

I am only speaking for myself now, but I believe that a clean break with 
version 3 is the best way to go. GNATS 4 is a very different beast from 
GNATS 3, and although it induces pain for sites that won't convert all 
their GNATS databases to version 4, it makes a very resource-starved 
development project much simpler.

Yngve Svendsen