But I don't like the second part where active users that have left get 'unpersoned', to use a George Orwell-ish term.

My best example of this is tilly, a fantastic contributor who has been forced by his employer to stop visiting the Monastery. (Professional Employees and Works for Hire) If tilly - or anyone - decides to return after a long absence, he should not have to start over as an initiate - his old posts should continue to bear his name.

s!!password!;y?sordid?binger?;
y.paw.mrk.;;print chr 0x5b ;;;
print;print chr(0x5b+0x2);;;;;
[download]

Well, as long as we go with the "No Write Up" part (and a looong time of not logging in), it should be fine. Even paco would get to stay then. :)

I don't think any user with writeups should be deleted, even if it wouldn't cause havoc in the database.

---

You have moved into a dark place.
It is pitch black. You are likely to be eaten by a grue.

It's easy to adjust the stats to account for inactive and zombie accounts. The wasted disk space in neglible. However the name space aspect of the idea is what gives it merit...wouldn't you hate to have your preferred nick taken by some zombie?

I think jeffa's remark about attempting to contact the person before hand is a neccessary step before feeding a zombie to the NodeReaper.

But this does present a problem though....a year old email address may have bit rotted and no longer be valid.

The email decay couldn't be avoided on the initial clean up, but going forward we could proactively email them one month after they register. "Hey, you've never used the account you've registered on PM. Do you want us to delete it?"

Perhaps they forgot. Perhaps they didn't like PM afterall. Perhaps we need a button that would allow someone who has never posted to terminate their own account...

/\/\averick
OmG! They killed tilly! You *bleep*!!

Hello,

ok - so to sum it all up:

• Something has to be done about it
• definition of a "zombie" is: loged in only once, homepage length 0, no writeups
• after n months of inactivity, an attempt is made to bring the zombie back to life with sending an email to him and explaining what will happen. If there is no reaction, m days later the user is purged. (*)
• No purge of nodes just because of inactivity (though I still think 2 years is long enough).
(*) While first being in favour of this, I can very well remember how I feel if I am "spammed" with email from some "uninteresting site". While I like the idea very much to try to bring back monks (not thought of in my post), we cannot assume everyone being as interested in this site as us.

Bye
 PetaMem

Most people won't take issue with a single polite email. After all, they did show enough interest in the site to provide the email address in the first place. It's not like PM wasn't given the address willingly. Sure there will be someone who will get bent out of shape about it, but the vast majority wouldn't. This really doesn't fit the definition of 'spam'...

In one year and a couple of months abigail will reach the two year mark. tilly will eventually reach it. Both of them have given a great deal of wisdom to the site. Node age isn't a metric of worth, and we already have a mechanism for removing worthless nodes. What we're really talking about is the possible removal of 'worthless' users/accounts. tilly and abigail may be gone, but their worth has more than been proven...

/\/\averick
OmG! They killed tilly! You *bleep*!!

Most of these Zombie Monks have no write-ups and homenodes of length 0. The top 10 largest homenodes collectively take up almost 300K. My two largest writeups are 19K and 15K. tilly's writeups total around 3 megs.

However, i do agree that at some point this might become an issue of relevence, but when that time comes i am sure that the committee of gods will best decide how to handle it.

jeffa

* subliminal typo - right before i submitted this my whole rack of CD's fell over! thanks for the catch Kanji ;)

Perhaps this can be a new superpower for the Saints who already have everything. Every 1000 points you get to take over a zombie monk.

But seriously, the only reason I can see keeping these Zombie accounts is if we were an evil company doing "permission marketting" BS and since we aren't we do the users a service by purging their information.

()-()
 \"/
  `                                                     

I'm not sure it would help, but after the registration page perhaps some pursuasive BOOKMARK ME NOW and ADD ME TO YOUR NAV BAR (all the most pious monks do!) might help get the numbers up a bit? I suspect people would return if they could remember where this page was!

In fact, I just stumbled on *my* bookmark today and returned after a long absence where I didn't remember this site was here.

tuxxe

Hi,
• They take up valuable name space
• They falsify the statistics (we're not a community of 16 thousand)
• They surely do have impact on search (because of the namespace)
But most of all: This "problem" will scale. In a year or so we'll be facing 10 thousand "Zombies" and so on and so on. We will have to do something about it, the only question is when.

Bye
 PetaMem

I think the name space is the most important of those you mentioned. Is it hard to choose a name for a new monk nowadays? If someone mighty could gather such statistics from Monastery logs, that would be interesting.

I don't thinks those zombies have significant impact on searches. And the statistics figures you render are not false, they're ... eh ... statistics. Anyone making bold clauses about the size of on-line community based on the total number of registered users lies. Meanwhile the wonderful Perl Monks Statistical Page already deliver workaround, as you yourself note in the original node :)

I can see a use for deleting true zombies. Maybe not a huge benefit, but namespace suck could become a problem.

However, loosing the author name of old nodes, because the user is passed is a no win. There is something to be gained for keeping an author's collected works. This provides continuity of opinion, shows growth of community members, and gives us cultural continuity. We are a community, and records of the past are important to the community. Besides, old nodes may point to an old user . . .so the link remains pointing to dead nodes? No, that won't do. Besides, it'd be a hassle to write such a tool, with dubious benefit.