[Mapserver-dev] Rewrites Considered Harmful

[Sean Gillies]

On Jan 15, 2004, at 2:52 PM, Frank Warmerdam wrote:

> Folks,
>
> I'm sure lots of you keep an eye on Slashdot and have seen this, or  
> have
> read discussions of the idea from other sources, but I thought it  
> might be
> worth forwarding anyways.  The artile is "Rewrites Considered  
> Harmful?".
> The basic thesis is that we should carefully consider before embarking  
> on
> wholesale rewrites of software products.  "Next Generation" software
> rewrites often cause far more pain then the gain they supposedly bring.
>
>   http://www.neilgunton.com/rewrites_harmful/
>
> I have personally seen this in a number of instances, including some
> rewrites I was responsible for myself.
>
> I bring this up here, because there is occationally suggestions made  
> that
> MapServer needs a complete rewrite.  There might well be some  
> substantial
> benefits to such a thing.  Lord knows parts of mapserver are pretty  
> frightening.
> But at the end of the day, I imagine we would find a rewrite would  
> introduce
> huge compatibility, bug and performance issues.  Furthermore, with an  
> open
> source project like mapserver, there is a very real danger we would  
> just run
> out of steam to complete the overhaul before it was really completed  
> properly.
>
> On a smaller scale, I also take the point that changes between  
> versions that
> break compatibililty have a higher cost than some might imagine.   
> Certainly
> I think that is true of mapserver, and backward incompatible changes  
> to the
> mapfile, or the various mapscript incarnations.  I was pleased to see  
> that
> proposals for the new "time indexes" did seem to try and account for a
> backward compatibility mechanism.
>
> Anyways, food for thought.
>
> Best regards,
>
> --  
> --------------------------------------- 
> +--------------------------------------
> I set the clouds in motion - turn up   | Frank Warmerdam,  
> warmerdam at pobox.com
> light and sound - activate the windows | http://pobox.com/~warmerdam
> and watch the world go round - Rush    | Geospatial Programmer for Rent
>

Hi Frank,

I really appreciate being reminded of these issues.  Is anyone surveying
MapServer users in a way that would let us know if we bungled the
version 3 to version 4 change?

I don't remember having seen any discussion of a complete rewrite, but  
then
I wonder if some of my ideas aren't perceived as being complete  
rewrites!
When I've written about my interests in seeing a thread-safe MapServer  
and
going away from the lexer towards thread-safe MapScript object  
marshaling,
do you see this as requiring a complete rewrite of MapServer?

cheers,
Sean

--
Sean Gillies
sgillies at frii dot com
http://users.frii.com/sgillies