[Moin-user] Why oh why!

Mike Fechner mike.fechner at consultingwerk.de
Thu Dec 11 03:28:54 EST 2008


> It's just my opinion, but it's important for serious software 
> to maintain backward compatibility at all cost.

from a professional working perspective, I can understand your point. But please don't forget: MoinMoin is an OpenSource initiative - it's free. Work of volunteers. Not a commercial software product you paid a license and probably product maintenance fees for. So the cost of OSS "maybe" that the user needs to be willing to spend more time on reading documentation before upgrading.

Even with commercial software (thinking of a big shop in Redmond) upgrades are a headache once in a while.

> Most users don't care if the engine is better. What they care about is, 
> they had something working

It has been said before: You could have sticked on the old version for ever.

I'm running 2 wikis at 1.5.something and a new one at 1.8. I also don't have the time to upgrade the old wikis right now - but I may do it at some time. Nobody forces me to do so.

(not involved in moin moin development at all, just a happy user since a while)

-----Ursprüngliche Nachricht-----
Von: Pierre Coupard [mailto:ppc at alum.com] 
Gesendet: Donnerstag, 11. Dezember 2008 09:12
An: moin-user at lists.sourceforge.net
Betreff: Re: [Moin-user] Why oh why!

Rick Vanderveer wrote:
> You cannot have a "legacy" switch because the whole point is to purge 
> that ugly code to begin with, if you kept it you'd only be tripling 
> the work for yourself!
I wasn't talking about keeping the old code around, but having parsers 
to convert the old code to the new engine on the fly, on pages that were 
made with the old engine. Perhaps even something as simple as having 
old-code pages transcoded to new-code pages that are cached 
transparently by the wiki whenever a page is edited: that way there's 
zero additional work to do, since the conversion scripts already exist, 
and the only performance hit is on the user when they save a page after 

It's just my opinion, but it's important for serious software to 
maintain backward compatibility at all cost. You can't expect people to 
put hours of work in learning something and building something with it, 
then pull the rug under them because "the engine is better". Most users 
don't care if the engine is better. What they care about is, they had 
something working (heck, even limping along, but working) and now they 
don't anymore, and they have to learn new things.

Remember: we're a firearms manufacture, not a software shop. In a 
computer-related outfits, people are used to put up with software 
antics, but in factories, they expect things to work. They don't 
understand that, no, it's not like a milling machine, it won't be the 
same for the rest of their lives. And I can also tell you that people 
who spend their lives mostly away from computers are very distraught 
when things change. It's just not part of how their brains work. The 
place I work at is reality, this is where things like MoinMoin get 
deployed. I have no right to tell you how things should be done, I know 
MoinMoin is fueled by people's free time and goodwill, but you should 
consider this also.
>   This is absolutely no different than any other software project, 
> whether you're talking about a small open source project, or a huge 
> operating system project. As we all know, every operating system 
> changes many behaviors between major releases (especially, it seems, 
> Microsoft.  But Apple or linux is no different).
> So, my advice is, if you want to maintain absolute legacy behavior-- 
> stay with that release, forever!  It won't expire on you, and assuming 
> it's behind a firewall it's perfectly safe!  A twenty year old Mac 
> Plus computer will still be able to perform the exact same tasks 
> (which isn't much, by today's standards) as the day it was bought.
Wow, way to mess up an analogy. Apple's MacOS is indeed notoriously bad 
at maintaining backward compatibility, because Apple wants to get rid of 
the old cruft. Microsoft however has an operating system that, however 
bad, still runs DOS software in emulation 30 years down the line, 
complete with the bugs of the time. This is the single strongest selling 
point of Windows. Microsoft puts a lot of effort into ensuring old 
software runs on their new OSes, and as much as I hate the company, it 
doesn't get praised enough for that. Guess who has the biggest marketshare?

I can't count the number of people I know who had a Mac and reluctantly 
ditched it to buy a beige box PC when a new, incompatible Mac came 
along, and they looked at having to buy new versions of their 
already-paid-for software to fit the new MacOS. They did it once, they 
didn't do it twice.

> Since you already bite into that apple, you might as well go the rest 
> of the way.  From 1.7 to 1.8, it's a baby-step.  Might as well get 
> fully current since your wiki is effectively disrupted anyway, and 
> take advantage of *all* the latest features!
I don't need to be current :) What I need is to get back to my real 
work. I have enough MoinMoin newness installed to try to implement what 
I need, and the newest and greatest bells and whistles are not worth 
pulling the rug under my users yet again.

Thomas Waldmann wrote:
 > Please DON'T post it to the list. It will only confuse people.

I don't have it anymore, problem solved.

 > And I can assure you that a AWK script is unlikely enough to convert 
your data correctly.

It was enough for the few monkey tricks we use here. When you wrote your 
scripts, you had the burden of maintaining the integrity of anybody's 
wiki archive, so you had to include all features. For me, a few regexes, 
some grepping and some manual editing did the trick because I only have 
to take care of one simple wiki.

SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you.  Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
Moin-user mailing list
Moin-user at lists.sourceforge.net

More information about the Moin-user mailing list