Mixnode

Common Crawl has updated and is now at “3.2 billion Web pages and 260 TiB of uncompressed content”. In September they added a list of university domains to the crawl. This time, for the first time, they’ve actively been trying to blacklist spam-network pages.

The Crawl is also now including 300m+ new URLs from a paid service called Mixnode, which looks like a very interesting on-demand custom-crawling and indexing service…

“Mixnode can breeze through thousands of URLs per second and download gigabytes of data per minute without a hitch.”

Presumably some of this comes via abstracting sections of the Common Crawl, then ‘filling in’ the rest?

Now all Mixnode needs is a half-decent ‘public search’ front-end for a Mixnode crawl, and it’s ‘Build Your Own Search Engine’ time — without the limitations of a Google CSE.

Facebook changes Group picture headers

Facebook has changed code and this has affected the size of the picture headers on all Facebook Groups, thus radically messing up people’s headers. For now, a new upload picture size of 820px by 384px will give a Group admin about the right size and proportions, but it’s still not ideal re: spacing and crispness. I’ll update this post when the new correct size is known.

Update: marketeers are suggesting 1920px as the ideal new Group cover picture size, for upload. That’s the new Kindle Fire HD 10″ size, so I’m guessing that’s perhaps why Facebook changed the size on Groups.

Added to JURN

Rethinking ecology (novel ideas in ecology)

Biosystems Diversity

Traditions (sports shooting)

Sci-hub loses domain-name

The sci-hub.io domain name has vanished from the DNS system. The Register has the details.

“Guide to academic search” – updated

JURN’s page “Guide to academic search” has been checked for linkrot and updated.

Desktop for Instagram

The unofficial “Desktop for Instagram” lets you post to Instagram from a desktop PC with the Chrome browser. At least, it lets you post for now. Previous such possibilities were stomped on, after a while.

You might want to be slightly more wary than usual on this. No news media have yet mentioned it (judging by a Google News search), it only has 20,000 users, and it uses the trade-marked Instagram name despite not being connected with the company.

oaFindr now called 1Findr

Nothing about this on Google News, but apparently the universities-only oaFindr has dropped the OA bit of the name and is now called 1Findr…

“1science has decided to change the name of the product line. Market analysis shows that its greatest value is the analysis and the integral access to the content, and to reflect this change, they have changed the prefix “oa” to “1”. So, oaFindr becomes 1Findr; oaFigr becomes 1Figr; and oaFoldr + becomes 1Foldr Data.”

Por un lado, 1science ha decidido cambiar el nombre de línea de productos. Puesto que su mayor valor es el análisis y el acceso integral al contenido, para reflejar este cambio, han cambiado el prefijo “oa” por “1”. Así pues, oaFindr se convierte en 1findr; oaFigr se convierte en 1figr; y oaFoldr + se convierte en 1foldr Data. … proporciona hipervínculos a 30 millones de artículos disponibles en acceso abierto.

Source, SEDEC Bulletin, Nov 2017.

The same source also quotes some current statistics. According to the CEO 1Findr currently … “provides hyperlinks to 30 million articles available in Open Access.” as part of a wider mixed database of records for 85 million. That’s both paywalled and OA academic articles. Apparently dating from the 17th century onwards, interestingly. In comparison, Microsoft Academic’s last known total (late 2016) was 140 million.

DuckDuckGo multi-columns fixed

DuckDuckGo Multi-columns is now fixed: DuckDuckGo – Multi-Columns v.10 at Userstyles.org. This gives you DuckDuckGo search results as a multicolumn layout, suitable for a widescreen desktop PC user rather than a tablet user.

If you need to tweak the rather garish default colours, either backup your existing tweaked script before updating and then paste your prior snippets of code back in again, or use my colour-tweaking guide to fixing the code. It’s only a five minute job.

Note that, to access script editing, you no longer go to User Scripts in Firefox, but rather to: top menu | Tools | Addons | the Firefox Extensions panel | Stylish | Click on the Stylish “Options” button | DuckDuckGo – Multi-Columns v.10 | Edit | Save.

Blooming OA

2 million pictures extracted from the Biodiversity Heritage Library: Search: bookcollectionbiodiversity | Flickr. Size and crispness is of course variable, due to the nature of the materials, but there are plenty of good 2MP-sized picture to be found.

Not currently well indexed by Google Images, so one can’t do a site: + “Larger than 4MP” search in a useful manner. Nor is Flickr’s own search very good, seemingly. For instance, from 2 million images Flickr can only find 7 for the keyword “flower”…

Element Hiding Helper in Firefox, how to get it back

Even if you are avoiding the new Firefox 57, beware of updating the new Adblock Plus browser add-on to 3.0. According to the developer, the vital Element Hiding Helper for Adblock Plus 1.4 is…

“discontinued due to the new Firefox extensions system.”

You can get back to 2.9.1 (June 2017) on Firefox 55 or 56 here, which has a 2.9.1 version that still works with the latest Element Hiding Helper.

Update: I gave up on Firefox and moved to Opera.