Author talk:Samuel Tolver Preston

From Wikisource
Latest comment: 1 month ago by Marjan Tomki SI in topic Works: Physics of the Ether (1875 )
Jump to navigation Jump to search

Works: Physics of the Ether (1875 )

[edit]

I found it available at Internet Archive on URL [1]; the contents is declared:

Book digitized by Google from the library of the University of Michigan and uploaded to the Internet Archive by user tpb.

Contents of the book opens in a viewer that allows searching, but results seem to be a bit quirky: if I search for Ether, a lot of hits for ether are shown, but the first match shown should be the one in the title, and that one was not shown. So I can't be sure all the hits would be shown (and to verify that would have to either go through it visually, or export it to another tool to search with I can trust more). I tried to search also for eTher and seem to have got the same results as with Ether (and the same count of hits, which I had to look about a bit to find), which means search is case insensitive (with possible exemptions of non English characters, which I didn't test).

So if I search for something to prove it is not mentioned in the book (like e.g. Faraday, and atom I tried and got answered as "0 results"), at the moment I can't be sure it really doesn't exist in the text. It may be that part of the text was not, or not quite successfully, OCR-ed after (as far as I see, good) initial scan, but I can't be sure about that either yet.

Anyway, link to the contents of this book might be useful in a section of external links, or maybe the contents (possibly scanned from the same University of Michigan source for rescan if need be) could be uploaded to wikisource too. Marjan Tomki SI (talk) 11:19, 30 May 2024 (UTC)Reply

I found why first ETHER, on internal title page (the book seem to have ben rebound: internal title page mentions publisher in UK, external front-page Library of the University of Michigan) was not found: it was spelled (and scanned) as E T H E R (with spaces between letters, so it was not recognized as the word, so search mechanism there works reasonably well.
It seems I might still miss mentions of words I search for if they were mistyped in original, or mis-OCR-ed. And it seems (at least some of text) is OCR-ed (Optical Character Recognized) on the fly, so LONDON was recognized (by search) on the internal title page, "(All rights reserved.)" (in small print and Italics) on the bottom of the page, or any word of it, was not. That page seems to be shown as a picture, from which only a part couldn't be selected, where on others you could a word, or a continuation of text through several lines. On other pages search also finds all italic text I tried.
Search also finds all occurrences of the statement (multiword string) you look for instead of the statement, if you don't enclose the string in "" s (implicit or). E.g. "two forms of energy" would find the string and show a bit of context around, without double quotes around all mentions of every word would be counted and shown.
Useful (I now seem to be able to rely reasonably on hits, both found and not found). But it's still unusual (at least for me; and this remark might help somebody else too). Marjan Tomki SI (talk) 14:21, 30 May 2024 (UTC)Reply