Re: gEDA-user: VeriWell now on SourceForge

[Hagen Sankowski <hsank@xxxxxxxxxxxxxxxxxxxx>]

Am 29.09.2005 23:40:22 schrieb(en) Samuel A. Falvo II:
> On 9/29/05, Hagen Sankowski <hsank@xxxxxxxxxxxxxxxxxxxx> wrote:
> > I agree. But your attribute 'abysmal' belongs to a lot of cores I've seen on
> > opencores.org both in VHDL and in Verilog. I think there are to less fellows to
> 
> This same argument can be applied also towards software on
> sourceforge.net too; but does that mean all hosted projects are
> "sub-optimal?" 

I can't do a quality statement to sourceforge.net, software isn't my business. Over the years I tried often to get in contact with some maintainer on opencores.org and failed. The guys are gone, the projects seems to be death or doesn't rising up 'cause the menpower is missing. I agree with Evan that a peer review process the way I know at wikipedia.org is important; the fellows dispute over all entries and couraged to erase bad ones. 

Well, how can I kick a maintainer to follow his own planes? Sending emails, offer my help, and still wait. But regarding the low complexity of the designs I often play with thought to fork 'cause I doing the job better. Patch a bad source is a mess, and sometimes I faster in rewriting the code than in waking up a death project or maintainer. I hate to waste my time. The one project I hesitate to patch 'cause of the bad source and I hesitate to rewrite also 'cause of the complexity is the PCI-WB-Bridge. The documentation is quite readable, but the source belongs to the group of "If it works, it's right" designs. I like the group of "It doesn't work unless it's right" designs (phrases friendly take from NetBSD).

How Linus said? "Do a good job"? "Release often"? Why this points are missing on opencores.org? Maybe there's a problem with the perceive quality. BTW, to perceive quality, look at the testbench.

> By the latter, I mean something where a
> hardware (or software's) design is so fundamentally flawed or
> out-dated that a grass-roots rewrite really *is* the shortest path to
> market.  In this case, then yes, go ahead and re-write.

I agree with you. In addition I would say there's a unvisible line of effort and complexity. Behind this line a rewrite isn't the shortest path. But a lot of opencores.org designs are still in front of the line. I think there are just a couple of heavy wight designs, isn't it? I wish, as a sign of quality, they becomes more.

Regards,
-- 
Hagen Sankowski              Email: hsank@xxxxxxxxxxxxxxxxxxxx