Index talk:Fairy tales from Hans Christian Andersen (Walker).djvu

Validation

@Cygnis insignis: Any special instructions for validation? Outlier59 (talk) 22:06, 21 January 2016 (UTC)

@Outlier59: Hi, nice to meet you. I had a look through the text layer to see if there was anything to look out for, but it seems it was pretty sound. There are more accurate proofreaders than me, so I hope there are not too many errors to distract you from what is one of the better translations (check out w:Jean Hersholt's at this site). My general approach is to transcribe a 'clean text' so replication of formatting is minimal (ignored unless it is arguably meaningful). Where there are subheadings within a story I notice that I decided on an extra line before it, like here, and hope that was consistently applied. You can use an edit summary to draw my attention to anything mystifying, or ping me again if I fail to notice that on my watchlist. Likewise, if you think something could be done better then let me know. Cheers, CYGNIS INSIGNIS 10:19, 22 January 2016 (UTC)

Oh, images. My view is that the reader should decide how big the served image is—and browser or user preferences be honoured—if they want to look at it any closer. I have seen |upright=value used when increasing the page size is justified, but haven't tested it and rarely thought it justifiable. This approach is the equivalent of using thumb in a wikipedia article. I have been tempted to assign a size to images in the past, but found the arguments against this persuasive; after all, a bigger image is a click away. CYGNIS INSIGNIS 11:56, 22 January 2016 (UTC)

Glad we got that sorted out right away. I'll leave the thumbnails as they are. Thank you! Outlier59 (talk) 12:03, 22 January 2016 (UTC)

I'll try to do one story every day or two. If you see me do anything you don't like, leave me a note here. So far, all I've come across are a couple of scannos. Outlier59 (talk) 12:56, 22 January 2016 (UTC)

I like that you are improving my transcription, cheers. CYGNIS INSIGNIS 16:37, 31 January 2016 (UTC)

Validating your transcription here is a pleasure, Cygnus Insignis. I see nothing but little minor stuff to change. :) Many thanks for checking my edit summary notes and answering sic questions. Cheers to you, too! Outlier59 (talk) 00:16, 1 February 2016 (UTC)

Weird bug using page tool "Clean up OCR" --costliest/costhest

I've been running the page tool "Clean up OCR" on these pages while validating. Came across a very strange "clean up" change on page 23, last paragraph. Cleaning up the OCR changed "costliest" to "costhest". I think it's correct now. I added the word costliest to another page just to check this, then ran "Clean up OCR" -- and the word was again changed to "costhest". I didn't save that change.

I don't know how common this sort of thing is when using "Clean up OCR". I'm using "Show Changes" before saving, so I think I'll spot OCR clean up glitches in these stories, because it looks as if "Clean up OCR" has already been run on the pages. It's easy to see the current changes before saving. It's almost impossible to spot clean up glitches if OCR clean up is left until validation.

I mention this because some of the errors I'm seeing (which are very few) might have been caused by running "Clean up OCR" after proofreading the page. Outlier59 (talk) 15:23, 23 January 2016 (UTC)

The script I use 'fixes' tlie to the, an error I happened to catch on that page, and it seems that "Clean up OCR" does the same thing. I searched on costhest to fix other pages where this has happened. There are outlier words containing this sequence of letters, thanks for identifying the problem Other59. CYGNIS INSIGNIS 20:08, 23 January 2016 (UTC)

That sort of automatic correction explains a problem I came across somewhere on another book sometime back. Wish I could remember the page. I couldn't figure out how to fix it, finally gave up. "Motlier" kept getting changed to "mother". But in that case I did not run OCR clean up after typing "motlier" over "mother" for the last time -- and it was still changed to "mother" when saved. Sigh. "Mother" is so much more common than "motlier" that it probably makes sense to assume "mother". At least "costliest" can be fixed! :) -- Outher59 Outlier59 (talk) 00:06, 24 January 2016 (UTC)

Pardon, thought I replied to this. I noticed that: mother, outhe, outher and another unfindable incorrection. These should be caught when someone reads through, though I miss some and others just click edit / cleanup / save and no more when validating. I don't know why it should keep correcting it, I use a older version of the same thing. You might suggest that the cleanup tool only corrects "_tlie". CYGNIS INSIGNIS 16:35, 31 January 2016 (UTC)

Edit counts

Between yesterday and today the edit change count on my contributions list has changed significantly for validating these pages. I have no idea why today's edits should appear about 140 per page, while yesterday's were about 14. Just want to mention I'm doing pretty much what I've been doing all along. Outlier59 (talk) 15:27, 31 January 2016 (UTC)

Sorry for the intrusion. Wondering if when you remove whole [empty] lines from the header/footer if it counts each [blank] space that is deleted from each line? A quick look showed that there may have been more line space removal in today's validating than yesterday. Londonjackbooks (talk) 15:48, 31 January 2016 (UTC)

Thanks for noting that, Outher, I can't explain why that would be. And thanks London for investigating the problem. I do glance at how many bytes difference occur as an imperfect indication of how many corrections were made, and use pop-ups to see what blunder I made. A clue might be that some pages were proofread years ago and others more recently, a probably useless references tag used to appear in the footer section and may disappear when it is edited again. Mysterious behaviour in the software is a frequent problem here, I used to make noise about this, but now find something else to do when this happens (and hope it has stabilised when I return). CYGNIS INSIGNIS 16:24, 31 January 2016 (UTC)

Okay, I won't worry about it. Thanks for checking it! Outlier59 (talk) 22:10, 31 January 2016 (UTC)