🚀 go-pugleaf

In article <1080@zinn.MV.COM> kgg@zinn.MV.COM (Kenn Goutal) writes:

[An excellent article on news & information flow in an organization]

>....
>I had previously written:
>>>Some careful planning of newsgroups and distribution regions
>>>should be done.  It almost certainly will not work to make all
>>>in-house newsgroups have a distribution of 'company-wide'.
>
>And Nick replied:
>>I disagree.  If we start distributing certain newsgroups just to 
>>certain people, we might as well (except for the minor issue of
>>resource usage) go back to using e-mail aliases. 
>
>First, I minor quibble:  I don't believe that resource usage is a
>minor issue for many, if not most, organizations.  

Clarification: for written electronic communications (not necessarily 
graphics, sound, video, or source code), resource usage is an insignificant 
fraction of labor costs.  For example, a 1 KB e-mail message 
to 1,000 people takes up at most $5 of disk space if everybody saves it 
in their folder, while taking 1,000/60 * $30 = $500 of labor if everybody 
takes 1 minute to read it.  The savings by going to news for some 
communications is due to the user interface news provides for filtering 
information, and the easier access to more information from getting rid of 
distribution lists.  The resource savings is small compared to the
labor savings.


>If it were, we'd
>all have disks on our workstations, maybe even big ones, and point-
>to-point connections with all other nodes.  We don't because resources
>are expensive.  

"All other nodes" (even more than a few hundred), without a good filtering
user interface, would quickly overwhelm the user.  Labor costs, the ability 
of people to communication with and via the computer, are a limiting 
factor.  There are organizations that have no need for terminals on every 
desk other than electronic communications; for these groups the issue is not 
news vs. mail; it is electronic vs. paper.


>So, I think that for e-news to be an effective tool, there must be
>some newsgroups whose distribution is org-wide (e.g. org.comp.languages,
>org.rec.aardvarks) and there must also be some newsgroups whose distro
>is limited to those machines where it is appropriate (org.BF1.hvac to
>discuss the air-conditioning problem at the Bedford plant), and there
>must be some whose distro is somehow limited to those *individuals*
>for whom it is appropriate (e.g. org.vp.strategy, org.grunts.strike).

org.vp.strategy and our.official.secret.club I can believe, but it
would take an awful lot of union influence in Sysadmin to pull off
org.grunts.strike.  :-)

What machines are appropriate for org.BF1.hvac and who decides?
Can a grunt working in a sweaty Bedford lab post the temperature?  
If I have health problems with certain chemicals and am traveling 
there from another site, can I post a question about air quality?  
Can a maintenence man at the Hartford plant who finds a similar
problem read the Bedford discussion to see if there was a solution?  I 
don't see how we can say in advance who needs and doesn't need to read 
a newsgroup.

Another example might be an "org.travel" newsgroup with good fares, 
hotel rates, etc. between company sites around the globe.  Some people 
need to read this often, some hardly at all.  The usual pattern will be 
"read it if you are about to travel."  This doesn't fit the concept of a 
distribution list, but some might be happy to shoehorn it into one.
"Execs and salesman who travel alot" would leave many part-time travelers 
out in the cold.  Just to pick another random example, information on 
competitor's benchmarks.  Which people or departments in the company need 
this information?  How can a list be devised that won't leave some people 
in the dark who need this information to do their job (better)?  Why even 
go to the trouble of making up such a list?

IMHO, distribution lists are a historical artifact from the era of paper 
communications, when they were necessary for routing.  The productivity 
improvements from removing this barrier could be vast.  Of course, since 
distribution lists are also used for political purposes, getting rid of 
them won't always be easy.  But the organization that does so will be able 
to communicate better, be more cohesive as a team, and will end up winning.
Just the opening up of communications between marketing, sales, and 
engineering could have a major impact.  Since news currently does not have 
distribution lists, it is a good vehicle for introducing this change into 
an organization.


>I think that it logically follows that many people will needlessly
>suffer from the lack of many of the benefits of e-news just because
>there are some people in power in certain orgs who cannot afford to
>allow a black-market infonomy that bypasses all their controls.
>
>On the other hand, if e-news could be made non-threatening,
>by (a) speeding up the existing communication channels a zillionfold,
>(b) allowing denser networks of ad-hoc communications where they are
>already allowed, and (c) still providing controls on information where
>that is currently desired -- then e-news stands a better chance of
>being allowed in the door, and many people in the org will benefit.

In the long run, because organizations are competitive, the communications 
channels will have to change to derive the full benefits of e-mail and 
e-news.  Company-wide e-mail has already opened up many new channels 
that were previously impractical.   My interest lies in changing both 
software *and* organizations to find a new optimum, not just changing 
the software to fit current organizations.  


>>Mimicking the paper trail and all 
>>its politics with electrons is a dreadfully inefficient way to go.
>
>I don't see how this is necessarily true.
>I see that it is true if the existing paper trail and its politics
>are inefficient (aside from the mechanics of paper).
>....
>I see no evidence that all 'structured' organizations are inherently
>inefficient.  Bypassing the structure is more efficient than working
>through it only when the structure is inefficient.

It is not structure that is inefficient, it is how the structure is
implemented.  A paper trail/signature structure is (relatively) very
slow; an electronic distribtuion-list fast, a reader-choice structure 
faster *and* more flexible.  The difference is how long the information
remains queued up (usually a function of how many queues require human
intervention), and whether that information gets to those whose 
productivity it improves.


-- 
Nick Szabo			szabo@sequent.com
"If you want oil, drill lots of wells" -- J. Paul Getty
The above opinions are my own and not related to those of any
organization I may be affiliated with.