Forum Moderators: open

Message Too Old, No Replies

Deep Linking in DMOZ

there is more than you think about

         

cornwall

1:41 pm on Feb 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I came across an interesting analysis on **** that lists multiple included URL's in Dmoz

There appears to be a much higher degree of multiple listings than one might think given the oft repeated mantra

"Deep-linking is the exception rather than the rule."

The essence of it is that 10,000 URL's are included more than 15 times in DMOZ (Data is from RDF dump dated Sept 22nd 2002.)

startup

3:17 pm on Feb 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



"There appears to be a much higher degree of multiple listings than one might think given the oft repeated mantra"
There is more than one "oft repeated mantra"s that are misleading. Deep linking just happens to be one of them. Take alook at how many sites are listed more than once under the same category.

Travel

3:37 pm on Feb 3, 2003 (gmt 0)

10+ Year Member



There are a few catagories where it is actually very common to have duplicate listings for a domain which probably accounts for quite a bit of that. Many local sites are also somewhere else in the directoriy. Many news type catagories have deep links. The home section (recipes especially) is also full of many duplicated. I saw one recently with about 500 different deep links.

WindSun

3:49 pm on Feb 3, 2003 (gmt 0)

10+ Year Member



Epinions has about 2700+ listings now showing up in the Google directory, up from 2300 or so last time I checked.

DMOZ search is as usual not accessible.

cornwall

4:04 pm on Feb 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



And...

37743 sites have 5 or more listings.

Before someone in DMOZ does the calulation that is about 1% of all DMOZ listings.

I guess it depends what you mean by "exception" and "rule" ;)

Dynamoo

4:22 pm on Feb 3, 2003 (gmt 0)

10+ Year Member



It depends if the deeplink adds value to the category it's in and isn't redundant (i.e. not duplicated by a main URL link in the same cat or higher cat).

This particular mantra is that the ODP is for the benefit of surfers and not webmasters.

Epinions - 2,098 entries in the Home/Consumer_Information/ category out of 5,175 - i.e. over 40%. Most links are from pages of about PR5 or so. Hmmmm.

John_Caius

4:28 pm on Feb 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Any site can have typically:

1 topical link in its particular field.
1 regional link for where its office is located.
1 World link for each language the site is translated into.

So if you want more ODP links then get translating your content!

If your significant unique content happens to cover very different areas, e.g. cookery and soccer then it is quite acceptable to have a topical link in each section.

In addition, sites offering significant unique content across a wide speciality field are often afforded significant numbers of deeplinks. The Health tree is a good example of this - the most authoritative sources on conditions and diseases are often the large database sites and they are quite properly given links in the major disease categories.

beebware

6:11 pm on Feb 3, 2003 (gmt 0)

10+ Year Member



And don't forget that not all "deep-links" are true deep-links. A good example is Geocities sites. Yeah, you could argue (and count) that they are all deeplinks of the main Geocities site, but ODP editors tend to see each of them as seperate sites.

cornwall

6:38 pm on Feb 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



And don't forget that not all "deep-links" are true deep-links.

beebware

I realise that, but out of the 37743 sites that have 5 or more listings, very few of them are in the Geocities class.

In other words, although Geocities leads the list with around 100,000 sites linked from DMOZ, it is only 1 out of the 37741 sites with 5 or more deep links.

I am not knocking deeplinks (in fact I am all for good deeplinks, and feel that DMOZ does tend to be perhaps too conservative with allowing them ;) )

What I am trying to air is how much more widespread they are than perhaps popular perception might be.

ettore

5:51 am on Feb 4, 2003 (gmt 0)

10+ Year Member



>> 37743 sites have 5 or more listings.
>> Before someone in DMOZ does the calulation that is about 1% of all DMOZ listings.
>> I guess it depends what you mean by "exception" and "rule"

Double-listing in both a topical and a regional category is common practice, as said multiple times. Add to this that several sites -- in particular those based in non-English Countries -- often have at least 2 or 3 language versions (common situation is local language + English), and these sites, too can be listed in the appropriate World/[Language] subtree/s both in a topical and a regional subcat, and you can easily see how 4 listings of the same site are quite a common situation.

Also consider that the figures you give account for both deeplinks (that is, site.com which has listings also for site.com/foo/blue-widgets and site.com/blah/anotherstuff ) and multiple listings (site.com listed in different categories, or site.com and site.com/fr/, site.com/it/, site.com/deu/ etc listed in the appropriate language categories).

While I would tend to consider 1% really an exception and not a rule, in this case this figure is not even the correct deeplinks count, being instead the deeplinks+multiple listings count.

Thus:
a) multiple listings = quite common (topical/regional and language versions topical/regional)
b) deeplinks = rather the exception than the rule, and only when a site is offering significant unique content across a wide speciality field in different sections/pages. We do tend to be somehow conservative here, since we want to list sites, not pages
c) Errors, mistakes, improper deeplinks/multiples in the same or in parent categories = they do exist, for a variety of reasons, but whereas they are often digged out and shown an example of how "the rules are broken" chez the ODP, they certainly account for a very, very small percentage of the number of sites listed in the directory. And we do like to remove them when found.

jeremy goodrich

6:00 am on Feb 4, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Google's got a few deeplinks. :) I think they should stay. So should all the others.

But then, I don't edit there.

If I did, perhaps I'd start where you are about to go, ettore. Are you going to go through and delete all those green ball links?

:) Have fun!

choster

7:14 am on Feb 4, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



A few sites will skew the totals. Geocities has been mentioned (and Angelfire and Tripod, implicitly). Some other outliers include

- news archives (accounting for ~40,000 listings)
- several dozen newspapers (the major US and UK broadsheets may have up to a thousand articles, features, or sections deeplinked)
- several hundred universities (both within the category for the university, and for research, publications, personal sites, or organizations listed elsewhere-- up to a thousand links for any major US research university)
- encylopediae (over 12,000 from the Catholic Encyclopedia alone)
- imdb.com (over 6000 listings)
- several hundred government and military sites (found in similar categories as university-sponsored sites; a thousand listings for nps.gov alone, over 2000 for navy.mil alone)
- several dozen e-text archives such as the UVa E-Library or the Perseus Project
- recipe sites (the category guidelines for recipes allow multiple deeplinks)
- genealogy sites (often editors may deeplink pages for a site devoted to several distinct surnames or branches)

This is far from saying that unnecessary deeplinks don't exist in the directory. But it does illustrate the kinds of sites which are deliberately deeplinked (as opposed to the majority of submssions, where deeplinks are occasional at best). If you think you see a case of abusive or unnecessary deeplinking, please report it to a meta-editor or at resource-zone.com for investigation.

rfgdxm1

7:25 am on Feb 4, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I am guessing these are mostly outlier cases like choster mentioned, or multi-language listings like ettore. And most of the rest are just editor error or inefficiency. Any abusive editor who tries adding lots of deeplinks for a site is just painting a big bull's eye on himself to be zapped. This kind of abuse is just to trivial to detect.

cornwall

8:11 am on Feb 4, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Any abusive editor who tries adding lots of deeplinks for a site is just painting a big bull's eye on himself to be zapped

That is the other side of the coin, I suspect that many "deserved" deeplinks are rejected by non-abusive editors who want a quiet life from their masters!

Getting the right balance is the key.

Dynamoo

10:23 pm on Feb 4, 2003 (gmt 0)

10+ Year Member



This raises the question of whether it is better to have focussed content sites with just one topic, or diverse content sites with several topics deeplinked in the directory?

rfgdxm1

2:15 am on Feb 5, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Skittish editors Dynamoo might be hesitant to add the deeplinks, so go for heavily focuses sites.

shangri_la

12:50 pm on Feb 5, 2003 (gmt 0)

10+ Year Member



I am an editor on Dmoz.
Currently there is a coloured tag system in use. Some URLs receive a green tag to actually encourage deeplinking. I believe these green tags are mostly used in the Arts categories. Green tagged sites can be band resources or other multi-topic sites, so long as they have good content. Sites might include rollingstone.com & other good resources with unique content; free website hosts (including free webspace that may come with ISP packages); MP3.com deeplinks are accepted if there is no official website... it's really up to an editor, and even if a site is tagged green, it is also noted that it must be reviewed to see if it is still worthwhile adding.

Multi-topic sites that are not deeplinked, are sites that have minimal content. For example, if you have a site made up of bands, with a brief biography, a common band photo, and a discography, and a few album reviews, with the sole purpose to push/sell affiliated links (i.e.: Amazon/cdnow etc...), then you will most likely not be deeplinked -- no matter how lovely/fast the design. That is not to say you cannot have a link to an affiliate on an album page. The key here is moderation, and an editor can pretty much tell what your motives are.

If you have a site that manufactures motorcycle helmets, and you sell those helmets online, you can:-
Have a listing in Shopping/.../Helmets, Plus a listing in Topical (i.e.: Recreation/.../Manufacturers or similar), or in Regional (your country/Locality) - not both Rec and regional.

Regional UK is the exception to the rule, as is multilingual sites. I don't edit in Regional UK, so I'm not going to comment on those reasons, but apparently they are allowed 2 listings in UK regional or something.

If you have a Shopping site that sells everything from women's clothing to Dog deodorant, don't expect multiple category listings in Shopping/.../ You will receive only one listing and it is up to you to write a brief (without the bells & whistles) description stating what you sell. If you add a physical address to your site, you might also be listed in Regional/your locality/Business and Economy/. But write a physical address, as I have seen Regional editors reject a site, if they only state "Located in the Houston area". You won't be listed in Houston unless you are in Houston.

The other day I searched Google for "Rhodesian Ridgebacks" (a dog breed), and the ad listing that came up was for pet insurance.

motsa

1:36 pm on Feb 5, 2003 (gmt 0)

10+ Year Member



Have a listing in Shopping/.../Helmets, Plus a listing in Topical (i.e.: Recreation/.../Manufacturers or similar), or in Regional (your country/Locality) - not both Rec and regional.

Not quite true. If you have a brick and mortar physical presence, you are entitled to a Regional listing regardless of where and how many times you're listed in Topical.

John_Caius

4:33 pm on Feb 5, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



What motsa said... :)

shangri_la

9:37 pm on Feb 5, 2003 (gmt 0)

10+ Year Member



OK! That statement should probably go in the FAqs for editors, as I have been editing for about three years now, and I didn't know that.

rfgdxm1

10:03 pm on Feb 5, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



As a general rule, if it is also a brick and mortar business the site qualifies for one Regional listing. Theoretically there is no limit to deeplinks listed elsewhere, but since they tend to be frowned upon typically the number that get listed for any site should be very small.

Laisha

10:17 pm on Feb 5, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Welcome, shangri_la!

OK! That statement should probably go in the FAqs for editors, as I have been editing for about three years now, and I didn't know that.

And there is what has been said repeatedly: Editors use different criteria. Not only do they do the subjective differently, but the word often just doesn't "get out" to the line editors.

Nowhere near 100% of the editors even read the forums in ODP, and even many those who do are overwhelmed by the sheer numbers of posts and details to keep up with.

This is where Zeal outshines ODP: all editors must know the guidelines before they begin editing, and the guidelines do not change frequently.

motsa

11:34 pm on Feb 5, 2003 (gmt 0)

10+ Year Member



OK! That statement should probably go in the FAqs for editors, as I have been editing for about three years now, and I didn't know that.

But there's nothing in the guidelines that says you can't be listed in both Recreation and Regional so I'm not sure where you got that in the first place (perhaps there's a flaw in the Shopping FAQs/guidelines). If your editing outside of Shopping is limited, then you really would have no need to know the specific guidelines for other branches and your lack of knowledge wouldn't necessarily be a sign of a flaw in the system. Not saying there isn't room for improvement (of course, there always is) but a link to a list of branch-specific guidelines is provided in the general guidelines so the information is available for those who look for it (note: I know the link is bad right now...that's been reported -- past versions of the guidelines included direct links to some of the branch-specific guidelines).

shangri_la

8:19 am on Feb 6, 2003 (gmt 0)

10+ Year Member



Motsa, Ahh... I see where you are going. I don't think I was clear enough in my first post. Anyway, I agree with rfgdxm1's comments.

Laisha, I do read the forums as often as I can. Sometimes I only have enough time to check-in every couple of weeks. Sometimes my questions are answered fast and clear. Other times there are so many different opinions, that the solution is as clear as mud.
I too applied to edit Zeal, and did so for a very brief time when it first began. The Dmoz guidelines may be vague in some places, however the reason for that, is to encourage more flexibility with future improvements. The few shortcomings with this system, are far out-weighed by the current and future benefits for the whole project. The current team of programmers are always improving things, particularly on editing side of things. It's actually a very pleasant experience most days. The improvements over the last few years have been outstanding to say the least. If the directory's image is promoted, and editors continue to volunteer to this great cause, it can without a doubt be the most invaluable project on the internet. The only drawbacks are lack of genuine active editors.

Certified Mozilla Fanatic!

cornwall

9:35 am on Feb 7, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Editors use different criteria. Not only do they do the subjective differently, but the word often just doesn't "get out" to the line editors.

Precisely, it is virtually impossible to have a consistent standard for multiple entries, as different editors have different subjective standards.

The overall effect is the mantra that "deeplinking is the exception rather than the rule" has led to junior editors in particular being discouraged from deeplinking.

Irrespective of whether the mantra is true or not in practice