Date: 02-09-10 11:46
Ok, I'd like to qualify the above and say that the book also mentions that if a scanned version is available, it can be downloaded freely. That's what I've done, and I can say I'm not disappointed with the pdf format, which shows a scanned copy of the book found at Cornell University, Ohio, USA.
> A Hakka Index to the Chinese-English Dictionary of Herbert A.
> Giles, and to the Syllabic Dictionary of Chinese of S. Wells
> Williams (Paperback)
> is reproduced via OCR method, but I would caution against
> buying it.
> Here's my review I wrote for Amazon. It hasn't been published
> by them yet, but you might know first hand what I've found
> having bought it.
> The errors inherent in character recognition in the OCR
> process is not a suitable method for this type of book. The
> original source is MacIver's index, but it is now a century
> old, and original copies would have discoloured so the
> quality of the text created from the OCR process may be low.
> This was the case. As an index of Chinese-English dictionary,
> it is completely useless. For anyone who has some knowledge
> in romanised Hakka, especially in the form that MacIver has
> published in, the numerous recognition errors made the
> resultant book we have on sale here like a string of random
> Here is a line of text from page 1 of the OCR reproduction
> f6747 j5Ll 578 Ion 7A"vi 5691 yiii nen V? L 298 (c. nsn.)
> 5809 311 khisn 7461 570 Ion 120t, 3248 ket J 447 khiefc
> Only "nen" and "ket" are romanised Hakka syllables, the
> others are gibberish. The rest of the book (75 pages) is of
> the same vein.
> I would have thought given the title of the work, that there
> was some sort of recognisable order in which the entries are
> made. Sadly this isn't so. As the text only reproduces text
> based on the roman alphabet, no Chinese characters are given,
> thus no context is given. The numbers do not inspire any
> recognisable order either. It is more of a book of random
> text featuring various combinations and permutations of
> random keyboard characters. In effect, it could have been
> typed by a child with a blindfold.
> My advice to readers is not to buy it. OCR is better suited
> to books where the text has context so that even with
> typographical errors, one can still make out some sort of
> sense from the OCR text. An Hakka Index to the
> Chinese-English Dictionary is not one of these.