[Pan-users] Re: Klibido and Pan (Was: Get new Headers?)

[Duncan]

Charles Kerr <address@hidden> posted
address@hidden, excerpted below, on  Sat, 30 Sep 2006
11:20:28 -0500:

> I played with klibido a little bit this week.  I like it and found much
> worth stealing. =)  I particularly liked the 'parts' column in the header
> window, that the task window showed completed tasks, and how you could see
> information about each part of a multipart.

Do you see now why I was having trouble explaining why it seemed to work
better?  It's a rather different concept, and until you see it in action,
it's really kind of hard to describe what it has that pan (presently)
lacks.

> klibido has a speed/memory advantage over Pan 0.115, though the advantage
> is much, much smaller than it was a week ago. =)
>  From what I can tell it's because klibido keeps most of its
> data stored in a berkeley database file, which has tradeoffs. We've kicked
> that Idea around for Pan before, and maybe that's something for the future
> (though I've had such horrible luck with Berkeley's stability that I'd
> rather use sqlite...)

Yes, a db /does/ have tradeoffs.  One of them is often stability vs speed,
unfortunately.  You probably noted the post here from someone mentioning
that klibido /was/ faster and better on memory, but corrupted its database
about once a week, so klibido certainly isn't immune from that tradeoff. 
Of course, the database corruption is the single most frequently cited
problem with bnr2 as well.  bnr3 is far more stable but far slower.  I've
not used either of them, but from what I've read from those that have,
bnr3 is slow enough it's not worth the extra stability, which probably
means it's worse than pan 0.114 was, /before/ the changes in 0.115.

> On the other hand klibido is a grabber, so it doesn't have anything you'd
> want for text newsgroups -- threading, charsets, quoting, scorefiles,
> GNKSA, etc.  That's not a bad thing, but a difference in focus from Pan,
> which tries to be very good in both worlds.

Exactly.  That's also the reason I'm now using pan for both, altho I like
some of the grabber (I use the words binary harvester) functionality of
klibido, and was using it for binaries before the new-pan series came out.

However, now that you've seen all those nice ideas, while I don't expect
to see them in pan 1.0, maybe for 1.1. =8^)  Even with no further
improvements in pan's memory and speed, if pan gets most of those GUI
features (and they don't end up slowing it down), I'll have no klibido
envy left at all. =8^)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman