Wikisource talk:Scan Lab

From Wikisource
Latest comment: 1 year ago by Mpaa in topic text at top of page
Jump to navigation Jump to search

wikidata and scan lab masterpieces and "instance of"

[edit]

I just removed a link to an ia scan from the wikidata entry that contains "our" djvu file. It was not the scan that was used in the djvu, for one reason. For another reason, the ia link was to a scan that was missing two pages. Another thing, it was not so great of an improvement in the images.

Images in scans
RGB scans are better than Grayscale which are better than Anti-aliased. RGB scans are usually with brown/pink/yellow paper. Grayscale is "black and white" but with smooth gray transitions between black and white. Anti-aliased scans have only black and white, so the non-straight lines on those scans are jagged, like stairs, and not smooth.

OCR likes anti-aliasing, good images come from RGB. Grayscale files are bigger than RGB (that was a surprise to me).

The DJVU file File:Grimm-Rackham.djvu was made from three different scans and some generated caption pages, for example. The like cannot be found online. Being homegrown, there is a good chance of spelling errors in the added "caption" pages, as I was the only proofer of that list and that is my weakness. It is of 3 different oclc works, 3 different oclc editions, 3 different ia, hathi links. It is a very very good wikimedia scan.

There are many "repaired scans" here, not just that one very awesome one. They should be able to say they were repaired in the data.

It would be nice to have "wikimedia scan" as an instance, which can be documented with where the different pages came from. Also a way to note scans that are not here as having one of the three qualities, (RGB, Gray, Anti-alias), if illustrated because, re-re-re-research is a drag.--RaboKarbakian (talk) 17:26, 4 October 2022 (UTC)Reply

@RaboKarbakian I think the most versatile solution would be to expand Proofread Page functionality to allow multiple scans (either complete scans, or scans of just the problematic pages) to be assigned to a single index page. I envision an option to toggle between different images of the same page as needs require during proofreading; that way, there would be no need to frankenstein together a single file from several sources. The provenance of each file could be maintained individually, instead of lumping everything together under one label. Realistically, we probably shouldn't expect such functionality to be added soon, but it seems like something worth requesting. Shells-shells (talk) 18:45, 4 October 2022 (UTC)Reply

text at top of page

[edit]

there's this "<templatestyles src="Template:Desktop only/styles.css">" text at the top of the page. i removed it from the source code editor thing, but it still shows up. can someone remove it? 96.241.104.204 22:23, 9 December 2022 (UTC)Reply

Should be OK now, there was a mistake in the template. Mpaa (talk) 15:34, 11 December 2022 (UTC)Reply