I saw a link to this on Resource-Zone a few years ago. It bugged me, so I wrote a counter arguement. Never posted it. I said I'd write more articles, so here you go, my response to Zvelo.com's article on DMOZ's demise. Context: DMOZ was a web directory edited by volunteers & owned by AOL. It had been online for nearly twenty years, but AOL decided to shut it dwon. I had volunteered to edit a less than a month before the shut down was announced. DMOZ's administrators were able to save the directory's data & reopen the directory under a new name: Curlie (https://curlie.org) If I sound a bit angry, remember DMOZ had just closed; I wasn't sure what would happen next, & that hurt a lot. I copied the portions of the article I responded to for context, without them this would make no sense, so my copying should qualify as fair use. I don't even get a link from this page; it's plain text. If anyone thinks this does not qualify as fair use I can remove this page. I'd recommend copying and pasting the URL below to go to zelvo.com. https://zvelo.com/top-8-reasons-dmoz-directory-web-dead-eol/ > Here are the top 8 reasons why we are kissing DMOZ goodbye today:" > > "No. 8: You snooze, you lose. > Updates to DMOZ were not made often enough." Does this refer to the layout, or the list of sites? As for the layout, it was updated in 2016 for the directory's 18th birthday. As for the list of sites, DMOZ removed dead links fairly quickly. Adding new sites took more time because editors had to sort through tons of spam, sites submitted to the wrong categories, mirror pages, etc. They also had to check links which had gone bad, & fix them if possible. The directory still grew over time, but in later years maintaining the links already in the directory took more time, so growth slowed. > > No. 7: Didn’t bring its bloodhound to sniff out the bad guys. > DMOZ provided no level of identifying malicious URLs – the bad stuff of the web. Malicious web pages could be a phishing web page asking for confidential information or download malware. Never saw a fishing page there. > > No. 6: A virtual boa constrictor slowly cut off DMOZ’s future blood supply. > DMOZ was restricted by base or host-level categorization – which never truly addressed path-level content. No. Think of when the directory was founded. In the late 90's most sites were hosted on a subdirectory of a ISPs server. At the time, DMOZ had to list path level content to have any content. And then how do you classify someone's homepage? They might have subpages for gardening, recipes, and Pokemon pictures their kids liked. Such a diverse set of interest had to be categorized at the path level, & it was. DMOZ also linked to news articles, encyclopedia articles, & scholarly articles when they suited the topic. > > No. 5: The arrow kept missing the target. > With no accountability for accuracy and little expectation for a correction – the user experience was not exactly a paramount concern. That's an insult to the editors who worked hard on the project. Editors tried to select a sites representing a diverse set of views on every topic. Note that style didn't factor into the selection process; information did. Still, listed sites were always usable. You could find information, good information, in DMOZ. > > No. 4: Left in the dust. > DMOZ did not keep pace with the times and continued to be inaccurate – including reporting a large number of false positives. More simply put: DMOZ became increasingly more irrelevant. There were few false positives in the directory. I'll admit the search engine needed work. To keep spammers out editors were discouraged from adding too many key words to site descriptions. The search engine worked, but I have a feeling a listing site keywords might've made its results better. > > No. 3: The Industrial Revolution is to manufacturing as artificial intelligence and machine learning are to automation. > DMOZ is a human-edited directory of the web and that’s just not a scalable business model for the size of the internet as it exists today. Newer technologies such as AI and ML can really turbo charge data analysis. No. I started using DMOZ because Google & Bing are bad at finding relevant results for general subjects. And they're getting worse at specific searches. They also have too few ways to refine your search. Clustering engines, like iSeek or Yippy, are better, but the categories they come up with aren't as intuitive as the categories in a good web directory. > > No. 2: You get what you pay for. > Free open-source is not always good enough. Historically known as the Open Directory Project (ODP), DMOZ was maintained by a global community of volunteer editors. Unpaid labor is not adequately incentivized to do a good job. That's an another insult to all the editors who worked hard on the project. In fact, it insults just about every volunteer effort. If you believe in a project you will do a good job, otherwise why volunteer? > > No. 1: Men stopped wearing pastel leisure suits with bell bottoms after the disco era ended. > DMOZ, an open source directory of the web, was conceived in the pre-Google era and never adapted to a post-Google (PageRank algorithm) search engine world. One of the primary use cases of DMOZ was to perform search engine ranking (which no longer operates the same way). Ok. A few errors here. First Google predates DMOZ. Google was founded in 1996 as a project by two Stanford students; DMOZ was founded two years later. Also, DMOZ was not designed as a tool for search engine rankings. It was designed to categorize websites so it's visitors could find what they need. Google's search algorithm used (and still uses) links to help weed out bad sites. Pages with more links pointing at them are generally considered better than pages with fewer links pointing to them. But it's not that simple. Google also considers the chance someone will click on a link & the chance someone will leave a page without clicking on any links when calculating & passing on PageRank. The earliest version of Google crawled Yahoo's directory to find new sites; when DMOZ came out it was probably easier to download DMOZ's rdf file to find new sites. Now they've been around for 20 years; they don't need to rely on someone else's list of sites as a starting point; they can crawl their own list of sites for new URLs. While DMOZ's main page may have had a high PageRank, it had many links; little of DMOZ's PageRank would've been passed on because the chance someone would click on any single link was low. It was probably better than other directories, because it was relatively spam free, & unlikely to be penalized, but it was not meant for site promotion. > The overarching reason for the demise or “DMIZE” (sorry I simply couldn’t resist) of DMOZ is the reality that it simply was not good enough compared with other more modern options that keep up with or even ahead of the needs of the web categorization market. Is any major player trying to classify webdocuments? All I see are Google and Bing, maybe Yandex. They just list sites; they don't classify them; they don't break them down into categories I can browse through. The only company I know of which tries to do classify sites automatically is Thunderstone, and they're not a major player at all.