View Single Post
Old 2012-03-26, 22:33   Link #527
Plorkyeran
Member
 
Join Date: Dec 2007
Quote:
Originally Posted by Zalis View Post
* Minor one: "Select Subtitles" no longer has a beep and or a dialogue box indicating "## lines were added/removed/set from/to selection." Not a huge deal since it still works the same, but it still would be nice to have that info.
It's now shown in the status bar since it's frequently not very important.

Quote:
Originally Posted by Zalis View Post
* Not-so-minor one: Spellcheck seems to have taken a step in the wrong direction. Mainly because it now parses hyphens (-) as if they were letters and not punctuation marks. (And for some reason, a standalone - always gets flagged by spellcheck in each new script, even if you clicked "Add to Dictionary" for it in another file.) Since typical fansubbed anime involves a lot of non-English words and suffixes, the Custom Dictionary comes in handy to store common honorifics, character names, foods, and other Japanese terms so that they don't come up as misspelled words every time. Pre-2.1.9, one could add names and honorifics such as chan/san/kun/sama to the CD, and spellcheck would recognize (for example) "Azusa" and whatever honorific as separate, valid words.

But now that hyphens get parsed, one has to separately add "Azusa-san," "Azusa-chan," "Azusa-sama," or other combinations, as well as all the variants like "Azusa-chan's", "Azusa-san'll", etc. Same with s-stuttered words: before, SC would see "W-What" as a standalone W (not flagged) and the valid word "What", and not flag it. But now, every new stuttered word shows up as a misspelling. The same also applies to words with -- at the ends of lines to indicate an abrupt cutoff. While "--" can be successfully added to the CD, SC will stop at a line like "I don't know why you--" and suggest that "you--" be "you --". Hyphenated compound words whose components are correctly spelled are also affected.

This change affects me a little more than average because I work with Magical Girl shows where the magical mascots end their lines with verbal tics like -coco or -mil. Pre 2.1.9, SC would not have flagged a line like "Look out-coco!" once "coco" was in the CD, but now every different [word]+[verbal tic] combination has to be manually added/ignored separately.

Overall, the dash-parsing in the new spellcheck takes more time, since depending on one's inherent accuracy, maybe 5-10 false-positive "hyphen issues" show up for every genuine mistake. And it may lead to more errors in final scripts, since people might blindly and impatiently hit "Ignore" or "Add" in [misspelled word]-[something else] or [misspelled word]-- situations, assuming that SC is showing an error because of the hyphen and not noticing the actual mistake. In essence, I haven't observed any benefit in the new way vs. the old way.

Anyway, if this has been brought up elsewhere in the past, I apologize for being redundant. And while I'm not one to criticize the time and effort spent in producing free software, I do hope that these issues can be looked into for future releases.
FUCK HYPHENS; HYPHENS SUCK; HYPHENS ARE DYING; HYPHENS ARE DEAD TO ME; HYPHENS HIT WTC

I don't think the removal of hyphens from the word separator list was an intentional change. There's no hyphens in the English dictionary we ship, so it's never actually useful to treat them as a letter.

FWIW 3.0 currently still treats them as word separators.

[e] Oh, apparently it was intentional

Last edited by Plorkyeran; 2012-03-26 at 23:25.
Plorkyeran is offline   Reply With Quote