Edit Weight

When I first submitted a boating site to the dmoz, I considered that the level of effort required for each edit to be relatively low (It's not I now assume).

Every single open boating category gave me the following "This category does not accept new editor applications at this time. It may be too large for a new editor, may be designed to hold only @-links, or may be reserved for more experienced editors for other reasons. Please try applying in a different category at dmoz.org."
I guess boating categories are not open to new editors.

The one boating editor has 3,970 sites in subcategories which have no editors (boating categories will not accept new editors). The average editor should have 60-70 edits overall attributable (3.8 million / 60, 000 editors). No one person can maintain 4000 sites.

Does any facility exists to determine editor weight and offer redistributed edits? It seems to me that dmoz suffers from data hotspots that some level of histogram could determine.

(A hotspot is a frequently updated or referenced area of data).

What is a possible solution?
1) A load balancing thread that allows an editor to tap for assistance from other editors.
2) A hot spot detection approach to determine if a bottleneck has occurred.
3) Use Industry assistance, advertise for non profit organizations that support commercial areas to help maintain these area.
4) Use less experienced editors, more often.

Allow me to try and answer your question from a general perspective, I'm sure someone else will give a more detailed answer.

I think that what is happening here is that you might just be seeing the tip of the iceburg and thus cannot see the entire picture. That's OK, though, and certainly no fault of your own.

We never, ever put a new editor in a large category or in categories that are "troubleome" because of abuse or spam. If we did, we would be setting them up to fail. If we don't allow brand new editors into these categories, it would be deceptive of use to allow people to ally to edit those categories.

It is probably worth noting that part of the iceburg you cannot see is that we have a different process for existing editors to apply for new category permissions, and there are different procedures for those applications.

While your thoughts on load distribution are interesting, I'm not sure that they would apply. Many, many editors edit at a level of comfort that they have reached. That may cosist of one small cat where there make an edit a month (and we welcome and cherish those edit and editors) or it maybe someone who has editing responsibilities in a dozen or more categories with thousands of sites and subcategories where they can edit. Editing permissions are not exclusive, so a person with editing permissions in a category that has 4,000 sites is not solely responsible for maintaining the part of the directory that has those sites. Anyone further up in the tree can assist in those cats, as can anyone with directory-wide editing permissions.

Also invisible to you are the mechanism we have in place to allow category editors to point out areas of the directory where assistance is needed, and mechanisms to receive temporary assistance.

Are there still backlogs? Yes.
Do we know where they are? Yes.
Are we doing anything about it? Yes.

So what is the problem? Why do they still exist? Well, the answer is both complex and simple at the same time. Let me give you the condense version: all of the editors are volunteers and, withing the scope of their permissions, can edit where and when they choose. A few hundred edits a month a are welcome; a singel edit a month is welcome. There are no quotas other then the minimum requirmeent of a single edit every four months in order to remain an editor. There are categories that editors shun: categories full of spammy, abusive and deceptive submissions top that list. These are categories where you dive into the pool of submissions, work for an hour or two and, if you are lucky, get to add maybe one site to the directory. It is very frustrating and discouraging work because editors do not feel as if they have accomplished anything. Oh, those cat do get worked, but not with the frequency of a cat maintained by someone in love with the subject matter.

Hope this give you some insight.

The entire Recreation/Boating is going through a major reorganization to bring it back to "Recreation". This is taking longer than I would have liked it to due to "life" and its many obstacles

I hope to be able to spend some time in the near future and have it completed by the end of the year. It is very time consuming moving sites to their regional categories one-by-one and many times creating categories and their charters/descriptions.

The apply to edit this category has been shut-off in many of these subcats due to the reorganization - hope this explains it

Is recategorizing common in dmoz?

Re-keying the categories, means a minimum of each category is moved by average depth/2 distance (say 3), so for 4000 sites you parse the cats 4000*3 = 12,000 decisions. Sounds hard.

In 24x7 databases, pruning a tree is easier in when one takes a subtree, reshapes the tree and reorgs it, then moves the branches to a new tree. Some of the branches are temporary placeholders for moving leaves to different branches.
Decisions are only made to group in a branch as a split then move, so decisions generally move 1 level down and then transfer or merge. That means the database stays somewhat balanced, and does not require an offline reorg.
(Those familiar with rebalancing AVT trees may recognize this as an insert/split reorganization.)

Is recategorizing common in dmoz?

Click to expand...

Mabe the answer could be "Yes". But you already figured out, why it seems to be common: A single large reorganization takes weeks or months (or sometimes years) to be done. So you will notice this reorg going on for quite some time.

See http://ch.dmoz.org/World/Deutsch/Computer/Internet/Webdesign_und_-entwicklung/Webdesign/ for an example. We started removing the letter categories more than a year ago, moving everything to the topical subcategories. Since only 3-4 people participate in this reorg, and all of them do some other things apart from that, it will take some additional time till the last few letters will be emptied (all reviewed and unreviewed listings have to be reclassified).

But this is a quite rare thing to do, so from another point of view the answer would be "No". To come back to the "How are most reorgs done":

Most reorganizations consist of simply moving categories (that are not in the best position) around or reclassificating the entries in small/medium subcategories one at a time. So most of them are much less work and easier to handle. From the user side you will hardly notice them as reorgs, more as single categories moving around.

What windharp said!

Also, time spent in a reorganization is not all consumed in the actual moves/changes. We are a concensus-based organization and place getting good concensus ahead of speed.

Since we are all volunteers, and a global enterprize, we cannot simply "call a meeting" to make a decision. We post, we discuss, we post some more, and we discuss some more. We may invite additional participants if a discussion goes in a slightly different direction, and so on.

While this would not necessarily work in a for-profit enterprize trying to meet marketing or sales deadlines, it actually works quite well here -- it just isn't fast (nor should it be IMO).

This particular re-org involves reviewing and moving 100's if not thousand of sites to their regional location. Many sites were added in years past that are shopping sites which do not belong in this branch. Category moves are great but when you have to review sites based upon their locality or what exactly they provide things slow way down.

Hope you understand

Editors can obtain more information about the re-org by visiting the thread in the internal Recreation forum.

While all the comments are cogent, since Boating is not accepting new sites to post, who is truly best served? Taking a large category "off line" to adding or update for months is the same as shutting a database off to users.

BTW : I'd offer to help, but boating sites are not accepting editors.

Where are new site suggestions not being taken? Only in the top level category of Recreation/Boating is the Suggest a URL link turned of, and that is normal for a top level category. Everything below that is still open to accepting suggested sites.

The reason that the Boating categories are not accepting new editors is because we thought it might be a problem to accept an editor for a category that might be going away in the reorganization process.

Please realize that major reorganizations are not unusual in the ODP. They happen on an ongoing basis. We never stop trying to improve the ontology of the directory to help users find the information that they may be looking for. We recently overhauled the whole Business branch and I'll bet you didn't even notice

Any subcategory of recreation Boating does not accept editors.

Not accepting new editors and not accepting new suggestions aren't the same thing. Yes, "Become an editor" is turned off. No, "Suggest a site" isn't turned off. So the category is still online to users and submitters.

Perhaps it would be a good idea for you to apply to edit in another section of the directory (not Boating). Then, when the reorg is done, you will have some practice editing and so you'll be able to apply to edit one of the Boating subcategories.

Turning off the "Become an Editor" link is not exactly disabling a database. Keep in mind that most of the ODP's visitors come to either look for a website or to submit their own website to the largest and most widely respected directory on the web. So, for nearly everyone, the database is entirely accessible. I'm sorry that the reorg has caused you an inconvenience, but remember that everything we do is aimed at helping our end users.

Hi,
Just posted this to see how the reorg is going.


I wrote the original posting and was not really surprised about the reorg not getting finished posted by the end of last year since it'll take about dedication of 66-200 man hours for any editors to reorg 4000 sites. 4000 sites times 1 minute per site = 66 hours, and those are calculated at optimistic levels. (More likely is a man month of effort (3 minutes per site) (shades of Fred Brooks)).

Hope things are going well.

Don't be discouraged! There are many women and some children editors.

Thanks LisaH, you sound like a number of Lisas I have known, - kind and sensible.

I am not discouraged, and while that may be true (women and children are editors), it's that editors like anyone need to have a degree of proportion in the edit jobs that they take on. No single person (or small group) can effectively change 1% of the dmoz in under 2 months without complete dedication to the task (what windharp said). One thing editors are is human.

Tilting at windmills (sorting out the 4000 sites) does few of us any real lasting service, and some of these jobs (OK, I am walking into the edge of blasphemy) are best left to semi automata. (A program recommends and a human approves).

Human editing for consistency of a computer database is not possible to maintain without tools to assist.

I'd do more, but I am not an editor, and like windharp says, we do best what we love. (I love boats).

Thanks


Regionalization is one task that the dmoz can leave to automata - especially if site submission capabilities are adhered to.