About tldlog

tldlog is an independent news log about the domain name industry. It lists the latest news, newest first, in two forms: our own short summaries of news from sources in several languages, each linked to its source, and original articles. It is published in English and Spanish.

It is written for domain investors, people who work at registries and registrars, resellers, and brand and intellectual property professionals. It also follows sources in other languages, such as registries in Spain, Germany, Latin America and Asia.

To follow tldlog in a feed reader, import the file with all our feeds (OPML).

Who makes tldlog

tldlog is an independent project run by Marc Gelabert, its editor and founder.

How pieces are chosen and reviewed

News links. Software reads the sources several times a day and groups the items that cover the same story. The editor goes through the list and chooses what to cover. For each chosen item, an artificial intelligence model drafts our headline, our summary and the “in plain words” summary, in English. The draft is written from the facts and never copies the source’s wording. The editor checks it against the source, corrects it and approves it. Nothing is published before that approval.

We never reuse a source’s images or paragraphs, and every summary names its source and links to it.

Articles. Articles are written and signed by people.

Sponsored pieces are paid for by advertisers, always labeled and not part of our editorial coverage. The advertising page explains how they are kept apart.

How our software reads sources

Every request our software sends to a source carries this user agent, the name it identifies itself with: tldlog/1.0 (+https://tldlog.com/about/). Requests come from Cloudflare’s network, not from a fixed network address of ours, and carry no cookies or credentials.

The software reads a source’s feed or, if it has none, its news page. For each item the editor chooses, it also reads the item’s page, once. It reads a source at most once every 30 minutes: usually every hour for a feed and every 8 hours for a news page. Its requests are conditional: if the feed has not changed since the last visit, the site can answer without sending it again, so an unchanged feed costs the site almost nothing. What we publish from what it reads is described in the section above.

The software follows robots.txt. The rules under User-agent: tldlog apply to it; if there are none, those under User-agent: * (the rules for all software) apply. To stop it from reading your site, add this to your robots.txt:

User-agent: tldlog
Disallow: /

To stop it from reading only your feed, put the feed’s path after Disallow: instead of /. If your robots.txt cannot be reached, the software reads nothing from your site. When a link redirects to another site, it checks that site’s robots.txt too.

You can also write to info@tldlog.com. To have a piece removed from tldlog, see “Takedown requests” below.

How we use artificial intelligence

Artificial intelligence helps with two tasks on tldlog: it proposes drafts and it translates. It never decides what is published.

• It proposes. For news links, a model drafts the headline and the summaries described above. When a source is in a language the editor does not read, it also writes a one-line description in English, so the editor can decide whether the story is worth covering.
• It translates. Every other language version is translated automatically from the approved version, never from the source, so the editor’s corrections reach every language. Software then checks that numbers, dates, names, domain extensions and links match the approved version.

Keyboard shortcuts

On the home page, the section pages and the other lists of pieces, a few keys help you move through the list. They need JavaScript, and they work only while a headline of the list has focus: press the Tab key until one has it. Everywhere else, the keys do what they always do.

• j: go to the next piece.
• k: go to the previous piece.
• Enter: open the piece.
• m: mark the piece as read, or as not read. The mark is kept in this browser only, like the pieces you open (see the privacy page).
• ?: show the list of keys. Escape closes it.

Corrections

To report a mistake, write to info@tldlog.com with the link to the page and what is wrong.

When an article is corrected after publication, it shows an “Updated on” note with the date and the reason for the change.

Takedown requests

A source that does not want one of its stories summarized and linked on tldlog can ask for the piece to be withdrawn.

1. How to ask. Write to info@tldlog.com with the link to the page on tldlog and the reason for the request.
2. What we do. We review the request. If it is justified, we withdraw the piece.
3. What a withdrawal means. A withdrawn piece is removed from every list on the site, and its page is replaced by a short notice saying that it has been withdrawn. The address still shows that notice, but answers with the code 404 (“not found”), so search engines remove the page from their results.

We follow 103 sources; in the last 30 days we found 0 items in them and published 0 pieces.