A platform for discussing a new web directory with a focus on human editing | Hacker News Reader

A platform for discussing a new web directory with a focus on human editing

curlz.org

24 pointssebst20 comments

DMOZ was great in the 90s (I was there.. and I remember using DMOZ and I appreciated it at the time). Search engines werent that great, and there werent that many websites.. 250k in mid-90s, and a few million at the end of the 90s. That's few enough that a group of people could review each one, and add it to a directory.

But now there are perhaps 250 million active websites[0]. Is it possible for a small group to review all of these?

And if it is, how many are directory worthy? I would guess there are enough high quality websites to overwhelm a directory...

My question is, with the amount of high quality content out there.. is a directory the way to find it today? And what about the near future.. how would a directory keep up with the rapid growth of the web?

The obvious answer here is to simply leave out websites.. which is exactly what DMOZ did.. it's been very difficult to impossible to get your website on DMOZ for many years.. but IMO, this is exactly what lead to the irrelevance of DMOZ, since it forced users to go elsewhere to find the content they wanted.

IDK.. I really like the idea.. but part of me also thinks the time for a directory has passed. (edit: at least in the form of dmoz.. maybe you can create a dmoz 2.0 that addresses some of those challenges though.)

0. http://www.internetlivestats.com/total-number-of-websites/ (this says 1B websites, but 75% are inactive)

I asked former editors of DMOZ if they would like to continue working on a human edited directory.

100 of them replied with mixed emotions, but nearly all of them are willing to contribute to the development of a DMOZ successor.

So, I build a platform (just a forum and a wiki at the moment) to collect all these likeminded people and start building something great together.

Tell me what you think about humanly curated web directories in 2017 or your ambitions to be a user or a contributor of a decent successor of the ODP.

Thank you for your feedback!

Starting from the other end, who would use a DMOZ successor?

Clearly DMOZ hasn't been getting a lot of real user traffic, which I think is because newer users (that's most users) have only ever used search, never been introduced to the concept of a directory.

So, the focus should be on who it's being built for, and how they would find and use it, rather than just replicating something from a different era.

For example, it could be a blog where guest bloggers (or blogger teams) each submit a carefully edited post describing the top sites (and perhaps apps) in a niche, and what's good about them. And then keep the posts up to date.

Quality could be maintained by a moderating team that only accepts the best posts, and ensures that they are maintained (otherwise deleted).

The resulting post when found in a search engine would be more readable, and probably more useful, than a directory listing page.

In my opinion you're 100% right!

Those challenges are real and a new version cannot be just a clone of dmoz.

AI/recommender systems could play a role in such a project, the relationship between discovery of new stuff and finding resources on already known topics has to be discussed and so on...

I invite you to bring in your ideas here or in the curlz.org forum.

humanly curated web directories

The word you're looking for is "portal."

Nah. The most common meaning of portal from when I last recall it being used was: "humanly curated selection of auto-updating RSS-fed content boxes with content from other sites".

DMOZ is a useful data source for what's on the web. Are there any machine-processable alternatives?

I use DMOZ to drive a list of major domains currently being used by phishing scams.[1] That's a list of who's in both PhishTank and DMOZ, ordered by the number of hits. (216 today. I used to try to get it down by nagging companies on the list, and got it down to 47 once.)

[1] http://sitetruth.com/reports/phishes.html

Someone mentioned the 4 yrs old blekko slashtag collection (https://github.com/blekko/slashtag-data)

I'm not aware of any other collection, but I hope curlz.org is gaining some importance soon ;)

I just signed up for "curlz.org". All they have right now is a forum, with a total of five posts. They don't even have a snapshot of the DMOZ directories. They need to at least host the DMOZ dump.

The entire content of DMOZ, in XML, is at

    http://rdf.dmoz.org/rdf/content.rdf.u8.gz

It's about 250MB. Save a copy, just in case. I just downloaded a copy.

Archive.org has a copy, at

    http://web.archive.org/web/*/http://rdf.dmoz.org/rdf/content.rdf.u8.gz

I will add the dump in the next days, of course

The Blekko slashtag collection is 4 years old, not updated, and only 5MB.

Thanks for the kind words, Animats. It's also a lot better than dmoz for a lot of search-engine-y purposes. Remember that the dmoz founding team is the blekko founding team.

The DMOZ data is licensed under creative commons and there is a RDF export. I downloaded the data, registered a domain - http://www.zedurl.com/ - and I'll try knock up a mirror of DMOZ tonight as a rails app. I'll update this post if things go well.

How about offering to collaborate with the OP?

There are endless link directories out there. Why do you think you can make a difference? Starting from fresh nobody will care to submit links because it's not really how SEO works anymore.

I doubt anyone just casually browsed DMOZ. A general link dir is probably a dead idea. A specialized niche one, not so much.

> There are endless link directories out there. Why do you think you can make a difference? Starting from fresh nobody will care to submit links because it's not really how SEO works anymore.

It never was how SEO worked, ever. The fact that DMOZ was a collection of high quality links made SEO jerks spam it to death but that was a side effect.

DMOZ was the polar opposite of search engines, though many of them took DMOZ as their starting point because of the high quality of the links.

DMOZ is how just about every search engine currently alive got seeded. The quality of the links was good enough that you could be reasonably sure to only hit spam after the second or third link traversal which gives a good testset to train a classifier with.

I've heard that point quite frequently.

So, one option for DMOZ to continue could be a movielens.org for bookmarks ;) But I'm sure there are further possibilities...

I don't think google, bing, or Yandex were seeded with dmoz.

I know you're probably just looking to replace DMOZ, and build something vast in scope/size. But please keep the idea I posted yesterday in mind:

https://news.ycombinator.com/item?id=13802555

Somewhat ironically/fitting one of the triggers for this idea was my dislike for the amount of corporate/institutional "spam" on dmoz.org. Browsing dmoz today it doesn't feel like browsing a treasure chest of information; it feels like browsing one of those vanity contact books that people are scammed into paying for placement in.

Just because 50 different museums from around the world are super reputable (to someone) that doesn't mean they should have a guaranteed place in the "art" category, for instance. They should need to have a kickass art collection online with great UX, super-highres images, easily downloadable if it's stuff that's out of copyright and so on. Just list the best ten (twenty? thirty? no idea what the right number is. it probably depends on the category) things, not every single art institution on earth that someone has bothered to submit.

I guess I would summarize this as opinionated, content quality-based selection.

Reply on news.ycombinator.com