Shrink the Google News headlines

Google News search results now have overly large ‘shouty’ headlines. Shrink them down a bit by adding this to whatever Stylus or UserScript CSS-fixer you have running…

.MBeuO {
font-size: 16px;}

Before and after:

Fixing Opera’s problem with Google Search + spellcheck

PROBLEM: The spellchecker in the Opera Web browser now only partly works in the Google Search box. Opera’s spellchecker underlines mis-spelled words, but can no longer offer any suggestions and make replacements. It works perfectly in other search-engines and on WordPress posts etc. But not in the Google Search box, for some unknown reason. Tests show that uBlock Origin is not preventing the spellcheck from working properly.

I tried a great many possible solutions, seeking a simple elegant right-click option that didn’t try to do 20 others things or send the word off to Whereizitagain to be ‘checked’. And which could work in Google Search.

SOLUTION: The Windows desktop offline solution I found was a combination of the old abandonware Spell Magic and the mouse-gesture freeware StrokesPlus (old version).

Spell Magic needs: General Options | Silent Spell Check Mode | On. While StrokesPlus needs: Preferences: Allow Left Click | On.

Then you set up your StokesPlus ‘S’ mouse-gesture with this mini-script…

USAGE: Once this is set up the user places their mouse-cursor anywhere over the mis-spelled word in the Google Search box. Then they right-click and draw the ‘S’ mouse-gesture with StrokesPlus. The gesture completes and the script then automatically runs the cursor back to the starting point of the gesture, and there does a double left-click. The word is thus selected and highlighted, and the script next calls Spell Magic into action on it.

The user manually chooses a replacement word, and then Spell Magic is able to make the replacement in the Google Search box. Spell Magic then exits to the Windows System Tray, ready for future action on some another word.

Also works in all other Windows-centric software, though sadly not on typed text in Photoshop. Spell Magic dictionaries are in the .ADM format and many are freely available from Addictive Software. One final thing to note is that Spell Magic will not launch if it thinks the word is correct.

On rare occasions Spell Magic will repeatedly fail to capture the word for checking. Exiting and re-starting the software seems to cure the problem.

Show Amazon Wishlist Total

A new free UserScript to show Amazon Wishlist Totals. Show the total cost of purchasing all items on a Wish List. Possibly useful for librarians and scholars doing a ‘bulk order’ via a finalised Amazon Wish List. Although that assumes you don’t need to bag the ‘used’ book bargains on the list, from third-parties.

OAmg tested

A quick test of the new OAmg academic search tool with keywords: Tolkien earendel. Four results.

1. “Authoring the Century: J.R.R. Tolkien, the Great War and Modernism.”

Book review, though initially looks like it might be the book itself. The Google Scholar pass-through button takes me to a result that then leads to a paywall. The English Society there invites me to purchase the item for $40. However OAmg’s PaperPanda button for the item takes me to Sci-hub, which instantly gives me a pirated copy of a review in the Oxford University journal English, 2010.

2. “Le livre des contes perdus.”

French translation of the Book of Lost Tales Vol 1. DOI button gives “error DOI not found”, as do all DOI links on OAmg, currently. “Download via Google” button is currently a template search unrelated to the search result. The Google Scholar button passes the title through to Scholar, which in this case is not that useful. You’re not going to find full-text for the French translation of the Book of Lost Tales via Scholar. There’s no PaperPanda link leading to Libgen.

3. “De Cynewulf a Tolkien: Earendel y las imagenes de la luz salvadora”.

2011 paper from an Argentine journal. DOI button gives “error DOI not found”, as before. “Download via Google” button is kaput, as before, though theoretically “a title as a phrase” pass-through to a Google search would be useful. The Google Scholar pass-through does get me the full-text link, on a university repository.

4. Tarscay’s “Chaoskampf , Salvation, and Dragons: Archetypes in Tolkien’s Earendel” which I know is free at Mythlore.

There’s a huge abstract on the OAmg page (each article/item gets one page), which is useful. DOI button gives “error DOI not found”, as before. The pass-through to Google Scholar gets the link to the full-text.

---

Not bad on basic numbers, then, on this micro-test. A 50% full-text delivery rate, and 75% if you count the pirated item. Other quick searches suggest that PaperPanda links (I’d never heard of them before) are all over the results, and the ones I tested all went to Sci-hub.

A large amount of open papers that ‘should be there’ for this Tolkien earendel search are not there. OAmg only claim a bare 200 million in total, via combining the outputs of all the usual API-offering aggregators.

A search for Mongolian folk song suggests the semantics is good, but also seems to indicate that the Chinese state aggregator is slipping in somehow, though CNKI or similar are not listed on the list of data sources used.

Still, full marks for chutzpah and energy. It feels like it could shape up to be a useful discovery tool, in the sense of “now I know this exists” rather than “now I know this exists, and I just downloaded the full-text”. But they have to: i) make the DOI pass-through work, which given the flakiness of DOIs is probably easier said than done; ii) hope that Google Search doesn’t throw a fit and block them when the title pass-throughs actually start working; and iii) throw Sci-hub piracy overboard (which seems as simple as un-plugging PaperPanda).

I’d say the value here lies in the possibility of timeliness and time-saving. Offer a $20-a-year custom monitoring service for new items of interest that happen to pop up across the 24 or so databases.