Talk:Long s/Archive 4

I removed the following chunk of cited text from the article.

===Optical character recognition (OCR)=== The long s is an example of the difficulties inherent in digitising old printed text; OCR is unable to differentiate between the long s and f characters.^[1]

It's not true; points out that Google was fixing this problem by 2013, Fraktur OCR programs have existed since 2005, Tesseract can apparently handle the long-s for some languages, I've hand-trained ABB&YY to recognize the long s, etc. Most current OCRs, designed for modern texts, have a lot of problems with the fonts and print quality of that era; the long s can be fixed by a semismart script (which could be integrated into the OCR program), whereas the c/e/o confusion in some fonts is a harder problem. I could clarify, but it seems to be making a big deal out of something that's just not that notable.--Prosfilaes (talk) 23:27, 3 April 2018 (UTC)

As you wish, but my experience with the British Newspaper Archive, where you have to search (for example) Suffex as well as Sussex, has been a frustrating one. Perhaps there is room for some note, as clearly some problems do still exist. People experiencing such problems might come to this article for an answer. I did. Tony Holkham (Talk) 23:38, 3 April 2018 (UTC)

Yes, but I imagine that the people who have to search for Russian expat newspapers in said archive have given up. That says nothing about OCR, that says worlds about the lack of interest of the maintainers of the British Newspaper Archive in handling anything but modern English newspapers.--Prosfilaes (talk) 23:22, 11 April 2018 (UTC)

I would favor including the sentence in some fashion, as many OCR programs still have significant difficulties distinguishing between "f" and "ſ," despite the advances that have been made over the past several years. This is still true, for example, on Google Books (certainly with Fraktur fonts), which is where the majority of people probably encounter the problem. What would you think of the following rewording? "The long s is an example of the difficulties inherent in digitising old printed text; many OCR programs struggle to differentiate between the long s and f characters." Bnng (talk) 03:49, 6 April 2018 (UTC)

If you want to digitize old printed text, get an OCR program designed to digitize old printed text. If you can't, then use text filters on the output as part of your complete OCR solution; using a standard dictionary to find all words that would be spelled with long-s, and converting the output words with f in them that should obviously be s and flagging those where it's unclear, as Distributed Proofreaders does, makes the long-s less problematic than the pre-modern fonts and generally bad print quality.

I won't fight a sentence that mentions the long-s as a problem in OCR, as long as it's not absolute, and preferably not under its own header, but I think it's less of a problem in real life; if you can search for Suffex, then you're better off than when you have several options for how the OCR could have mutilated the æ in the Latin text in the same work.--Prosfilaes (talk) 23:22, 11 April 2018 (UTC)

I think you're quite right, that it shouldn't be mentioned as an absolute. That is, there should be more context. The Traub et al. paper that is cited in the original blurb doesn't say anything approaching OCR being "unable" to differentiate the symbol. On page 256, they cite it as a "common error" that there is ample research into resolving (unfortunately, not citing any specific study or paper), and on 257 they note that the specific database they were studying noted that it would have errors that might include the long-s confusion. In fact, on 258 it becomes clear that the notation they discuss on 257 was a generic disclaimer; that the database maintainers had never formally evaluated the precision/recall rates.

Honestly, I think the Traub et al. citation ought to be ditched for something that discusses the background better. The blogspot blog mentioned above stands out as helpful, but there's a pretty significant resistance towards using blogs as sources, particularly in a field where there are academic publications. It does look like a version of that blog post was reprinted in the TUGboat publication, though it's a version from 2011. In any event, I don't think that the OCR issue is particularly appropriate for a freestanding section. The use of many short sections on very specific aspects of a subject is something of an artifact of an earlier generation of Wikipedia writing. Ideally OCR issues would appear alongside the Unicode table, whether in the "modern usage" section or in a separate section on computational issues or typesetting. —/Mendaliv/^2¢/_Δ's/ 18:01, 22 April 2018 (UTC)

In newspapers of the 1830's, I've seen a script p used to mean per ; and strike-through lb used to mean pounds weight. Where are these characters documented, do they exist in Unicode, and are there any other similar ones ? 78.144.83.200 (talk) 16:28, 20 October 2018 (UTC)

The Unicode characters are U+214C ⅌ and U+2114 ℔ respectively. BabelStone (talk) 16:36, 20 October 2018 (UTC)

Thank you, BUT " This site can’t provide a secure connection

codepoints.net uses an unsupported protocol."

78.144.83.200 (talk) 16:42, 20 October 2018 (UTC)

Although rare, capital form of long S exist, e.g.

At the height of the Antiqua–Fraktur dispute, in 1909, Fritz Helmuth Ehmcke designed an Antiqua font intended to preserve the advantages of Fraktur for the purposes of typesetting German without resorting to a bastardization of the two traditions. The resulting Ehmcke-Antiqua was unique in not only offering a capital variant of ß, but also a capital variant of long s (ſ) as well as capital ligatures of the ch and ck digraphs

https://www.typografie.info/3/artikel.htm/wissen/ehmcke-antiqua/

Can be other instances, too

217.118.83.207 (talk) 09:00, 16 June 2019 (UTC)

The present tense needs to be used in this section, because this is being done in reliable sources and in the source cited in this case. I believe the thinking is that when you pick up an old book from an earlier century in which the long “s” is used — the rules still apply inside that book. I think this is a helpful section, but it needs to indicate that the rules apply to the 17th and 18th centuries. The change to past tense was done in good faith, and it can be discussed here. I will restore the present tense and then add the centuries. - Bitwixen (talk) 12:51, 6 December 2019 (UTC)

And then it proceeds to tell you when to use the normal round S. This is really confusing. It should be explained the opposite way and explicitly state when to use the long S and not when to use the round S/not use the long S. Also, I don't get the remark about ſs looking like German ß. German ß is historically made up of ſs.--82.207.237.118 (talk) 08:27, 21 November 2020 (UTC)

Thank you. I have tightened the text a bit more but your comment is not really clear enough (or, more honestly, I don't know enough about the subject to understand it because my interest is typpgraphy) for me to do more. Would you please write downs here (or in the article) what you believe it should say, bearing in mind that whatever you write has to be supported by the only source we have at the moment, Babelstone blog: The Rules for Long S. Of course you may add new material but you have to back it up with a new source.

Wikipedia is entirely written by volunteers like you and me, so welcome to the project! I hope you will get an account but it is not required. Like many others, --John Maynard Friedman (talk) 12:46, 21 November 2020 (UTC)

I am not knowledgeable enough to clear this up myself, but the claim that this only applies to published English works of the 1600s and 1700s can't be correct since This New Testament translation from 1589 fits the bill. However, I can only cite an example, whereas the current article text here is footnoted and I'm not going to be able to properly dispute the claim or rewrite it accurately. 06:37, 27 January 2021 (EST) — Preceding unsigned comment added by 198.52.145.5 (talk)

Thank you for this valuable contribution. The Wikipedia policy wp:no original research doesn't allow editors to contradict the source in the body. The problem is how to identify when "custom and practice" became "good practice" and finally a formal rule. It may be that this edition of the NT was in the "custom and practice" phase—how do we tell? That is why we rely on professional sources. Perhaps you might write to Babelstone about it? --John Maynard Friedman (talk) 12:10, 27 January 2021 (UTC)

Regarding the bullet point: "Before a hyphen at the end of the line a long s must be used: Shaftſ- bury."

I have never seen the word Shaftesbury written without an 'e', neither when referring to the town in Dorset nor to Anthony Ashley Cooper and other Earls of Shaftesbury. Should there be an 'e' in Shaftſ- bury?Haynesta (talk) — Preceding undated comment added 12:31, 14 January 2021 (UTC)

Going back to the citation (which is not ideal, per wp:BLOGS, but it does cite its own sources), we find "short s is used before the letter 'b' in books published during the 17th century and the first half of the 18th century (e.g. husband, Shaftsbury)," which would suggest that the spelling has mutated since then. Unfortunately, the relevant volume of Victoria History of the Counties of England for Dorset is not online, as that probably gives the various spellings down through the centuries. If it bothers you, you could add a footnote? --John Maynard Friedman (talk) 13:12, 14 January 2021 (UTC)

@Haynesta:, I have just added a citation from 1705, written by the Earl himself at the time, that spells it without the 'e'. But I guess that that was a bit before your time. :-D --John Maynard Friedman (talk) 12:10, 25 February 2021 (UTC)

Thank you so much for this @John Maynard Friedman:. I find it so interesting to see and learn how the written and spoken language has changed through the years. 25 February 2021 Haynesta (talk)

I added, and Nardog deleted, this hatnote:

For the notation ⟨ ⟩, / / and [ ] used in this article, see grapheme, phoneme and International Phonetic Alphabet respectively.

so this is the WP:BRD discussion.

Is it reasonable to assume that all readers of the article will be familiar with these notations? The 'contains IPA' was an essential addition but is it enough? It would be great if there were an equivalent box for graphemes etc. (Is there?) John Maynard Friedman (talk) 14:47, 25 February 2021 (UTC)

The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion:

• Guinness & Co. - Guinness Extra Stout 2020.jpg

Participate in the deletion discussion at the nomination page. —Community Tech bot (talk) 09:37, 6 October 2022 (UTC)

Is there any reſearch on why long s was uſed? It uſed to annoy me, but I've grown to like it. I just haven't ſeen much in the way of explaining it. Would be a good ſection to add if we could find any ſources. Metallurgist (talk) 13:50, 14 January 2023 (UTC)

This blog at Grammarly good enough? Otherwise you have just volunteered to find an RS . I hope you know about Wikipedia:The Wikipedia Library? --𝕁𝕄𝔽 (talk) 16:24, 14 January 2023 (UTC)

I have moved a contribution by user:125.237.25.76, from the article to here at the talk page, because it is uncited and questionably of value since it is of long tail significance

• Sheets of Penny Blacks and Two Penny Blues from 1840 still have the long s in them as they say on the margins - "Labels must be affixed above the addrefs"
• Numerous political and satirical cartoons of the 1830s and even 1840s have long s in them.
• Framed prints of Queen Victoria's coronation in 1838 also still have long s in the words Duchefs and Highnefs, usually in Copperplate
• I have a envelope from 1905 that still has a long s on it. It was written by a 98 year old woman.

If anybody wants to pursue these, they will first need to find wp:reliable sources to attest them. --𝕁𝕄𝔽 (talk) 11:53, 28 November 2023 (UTC)