discuss: Software-RAID-HOWTO


Previous by date: 19 May 2008 08:59:04 +0100 Re: Software-RAID-HOWTO, jdd
Next by date: 19 May 2008 08:59:04 +0100 Re: LDP Process (was: Software-RAID-HOWTO), dominussuus.gmail.com
Previous in thread: 19 May 2008 08:59:04 +0100 Re: Software-RAID-HOWTO, jdd
Next in thread: 19 May 2008 08:59:04 +0100 Re: Software-RAID-HOWTO, David Lawyer

Subject: Re: [discuss] Software-RAID-HOWTO
From: David Greaves ####@####.####
Date: 19 May 2008 08:59:04 +0100
Message-Id: <48313417.3080200@dgreaves.com>

Karl O. Pinc wrote:
> This is a recurrent notion.  To summarize, without endorsement
> either way, the problem seems to be that tldp wants to
> be authoritative, in that content is should be vetted for
> accuracy, writing clarity etc.

And yet, when the author of the linux raid tools (ie the guy who writes the
kernel and userspace code, not the HOWTO) links to the community wiki and
removes the link to the HOWTO from the man page as it's no longer accurate or
authoritative - why does the LDP keep publishing it? (mdadm 2.6.6)

Well, we know. Inertia and resource constraints.
(not an attack - fact; good intentions with resource issues)

LDP is supposed to publish authoritative, accurate info.
It should *care* that the info it publishes is wrong and either just stop
publishing or link to a better source.

Has anyone looked at the dates on most of the LDP docs?
Can you, hand on heart, say that you are doing the community a service by
publishing documents about linux from 1997, 1998, 2001 even 2004 and *claiming
they are relevant*?

After a certain amount of time, documents should be reviewed and moved from the
display to the archives. The display (primary url) should say that the document
is considered out of date and there may be more relevant documents available
elsewhere. This may include submitted URLs.
It should, of course, point to the archive and it should invite readers to
participate and update the document that they are interested in reading.
This should be a manual decision but based on regular (annual?) reviews.

>  Significantly, this is a problem
> that wiki's have not solved.  At any given point
> any wiki page can be useless or worse.
Agreed.
Whereas a text document is guaranteed to be correct?

No. The difference is in the vetting process.
But, something else that wasn't forseen back in the day - maintenance and
deprecation.
And honestly - that's where a wiki can beat your existing process.
It lowers barrier-to-entry. A user spots an error and fixes it. The editorial
*team* has lots of small vets to do.
The existing process requires too many people to vet an entire multi-page
document. Many documents don't have reachable authors (I was lucky).

Checking a document for accuracy when it's published is all very well. And for
some subjects that's OK - facts tend not to change.
But surely you can recognise that LDP is documenting a changing world where
'facts' do change; the HOWTOs need periodic review and re-validation.

A wiki can provide accurate information and it can be structured to support
document management processes.
So can many other technologies.

> Put these two notions together, mix in a bit of
> not enough time spent on _any_ tldp content, and you get the
> present impasse where wikis don't seem to fit with tldp.
I think an ad-hoc process (as per many wikis) doesn't fit.
I think a technical mechanism allowing periodic updating of documents, change
history, authorisation and community inclusion may be worth considering.

> Tldp does not seem to feel that it needs to be the place
> people go for wiki content.  Other sites serve that need.
I noticed :)
I went elsewhere.
I am not waiting or asking for LDP to wiki-ise.
I am asking LDP to please stop publishing misinformation and suggesting they
link to a more authoritative work managed elsewhere :)

I am prepared to help.
I am not prepared to support the current process. It's broken.

> At least this is my impression as a long-time tldp list
> lurker, which I present so I don't have to read the same
> old arguments yet again.
I am not arguing that the LDP *should* do wikis.

I am arguing that LDP should:
* first and foremost : stop publishing HOWTOs when an authority (the new
maintainer) gives you notice that it's deprecated.
* open up to the community
* lower barriers to entry


> Frankly, somebody actually doing enough of _anything_ will
> probably get tldp moving and then inertia will keep
> it moving in whatever direction it's going.  Until
> somebody does something the present situation will
> continue.  Policy will be set by those who do the
> work.

As some staff have said and the recent changes pages show - no-one is doing much
of anything.

And clearly, continuing on the current downwards trajectory is not optimal.

David

Previous by date: 19 May 2008 08:59:04 +0100 Re: Software-RAID-HOWTO, jdd
Next by date: 19 May 2008 08:59:04 +0100 Re: LDP Process (was: Software-RAID-HOWTO), dominussuus.gmail.com
Previous in thread: 19 May 2008 08:59:04 +0100 Re: Software-RAID-HOWTO, jdd
Next in thread: 19 May 2008 08:59:04 +0100 Re: Software-RAID-HOWTO, David Lawyer


  ©The Linux Documentation Project, 2014. Listserver maintained by dr Serge Victor on ibiblio.org servers. See current spam statz.