[chef] Re: Migrating mailing list

[Lamont Granquist < >]

I sent this to Nathen and he requested I forward this on to the mailing list:

On 9/3/13 1:18 PM, Nathen Harvey wrote:
> We've looked at using Google Groups to host our fancy new mailing lists
> but that doesn't seam feasible primarily because it's going to be nearly
> impossible to cleanly migrate all of our existing messages to Google
> Groups.
Does that have to be a requirement?

If the old mailing list archives are kept around and searchable via
google search on the web the information can still be found.  As time
goes on the utility of that old information decays and will eventually
become more of a hindrance than its worth anyway.  I don't see why we
need a pack-rat mentality towards information here.

...

To expand a little bit by analogy: I solved my problem with having 20 
years worth of accumulated cables in buckets in my apartment by keeping 
the ones that were being used and recycling all the rest, and I was 
pretty brutal about it and only kept some spare ethernet and USB cables 
and tossed most everything else. Of course a week later I needed an 
adapter that I had just gotten rid of so I had to re-buy it on Amazon, 
but I haven't missed a cable since then and I've got significantly less 
clutter and all my 9-pin serial cables, old SCSI ribbon connectors and 
RGB composite video connectors are no longer cluttering my apartment (if 
I really need a 9-pin serial cable I can go down to re-PC and buy one of 
my old ones back...  and my ego doesn't need to center around being the 
guy who can come up with one at a moments nice...)

Recently, I noticed that if you typed "install chef" into google the top 
hit on search was for the old instructions on installing chef 10.x open 
source server. I also found that people were still blogging about how 
chef was difficult to install and citing that page as evidence of it. 
That came about a year after doing all the work to create omnibus and 
make the chef installation process much easier, and hanging onto that 
old information might have been keeping a few people who still use 10.x 
server happy, but was doing so at the expense of every single new user 
of chef out there.  That is information that is net doing substantial 
harm rather than being helpful.

You can see this on google when you find people googling "<subject> 
2013" to try to find recent information, because blog posts on ruby, or 
whatever, from 2007 often aren't at all relevant anymore.  I want to 
know about this years segfaulting bug, not one from a decade ago.

So, if we just froze the mailing list it'll be a bit harder for awhile 
to find useful information from the few months prior to the cutover -- 
but two years later all those old messages are going to mostly be 
useless clutter and misinformation anyway.  There may be a few gems left 
at that point, but the bulk of the information is not going to be that 
useful.  I don't necessarily suggest deleting the old mailing list 
archives, but I don't think its wise to make decisions about its 
replacement based on the requirement to preserve all the data indefinitely.

There seems to be an implicit assumption that the value of information 
is always positive, even though it may approach zero and that all 
information should be kept and preserved if at all possible.  I actually 
see information that isn't curated as crossing the zero point and 
becoming of negative value over time and that negative number constantly 
getting larger and larger over time as it clutters up search results and 
makes old wikis useless until they are ultimately abandonded because 
nobody can take a weed-whacker to them and delete old information ("but 
you might need it!").

OK, End rant...