Smartening punctuation

[roger64]

Hi

It looks very tempting but I did not dare yet to use it because I am French and we have our own -strict- punctuation rules.

Is there a way to customize this function to abide by our own needs or, at least, to check what exactly is performed on our files and if this is convenient for us?

[eschwartz]

Just like any global change it will create a checkpoint and you can use the differences feature. IIRC it will also automatically offer you in a popup to "see what changed".

Then you can revert if you need.

calibre as an ebook version control system is pretty handy, evn if it is limited to one branch and session-only.

[roger64]

Indeed the control point system is very handy. But, I can't help thinking I would feel better if I knew exactly what it is doing to smarten the punctuation in my language instead of having to check afterwards what has been corrected or broken.

[PeterT]

Just make a backup and experiment.

If it doesn't work properly, nothing lost but you could provide useful feedback on what is broken.

Do you actually expect that an Indian developer, speaking English would know if your arcane language rules are handled correctly?

[DoctorOhh]

Originally Posted by roger64

Indeed the control point system is very handy. But, I can't help thinking I would feel better if I knew exactly what it is doing to smarten the punctuation in my language instead of having to check afterwards what has been corrected or broken.

Review the info on this site. You may find clarification.

[roger64]

Originally Posted by PeterT

Do you actually expect that an Indian developer, speaking English would know if your arcane language rules are handled correctly?

I just expect to understand how "Smarten punctuation" works, according to which transposition rules. It may sound unreasonable to some... and if French looks "arcane" to you, in Toronto you probably may find help not too far...

@DoctorOhh
I'll have a look. Thanks.

[DiapDealer]

I understand not wanting to smarten a whole book as a test. Backups or not. You couldn't possibly know what might have gone wrong without then proofreading the entire book for potential problems. No fun.

I recommend making a test ePub that consists of only one page. On that page, throw every possible scenario you can think of (a paragraph or even a sentence at a time), "smarten" it and check the results.

Repeat until you are satisfied you know what's going on.

[eschwartz]

@DiapDealer,
But the whole point of my suggestion is using the diff feature to tell what changed, so there is no need to proofread.

[roger64]

Originally Posted by eschwartz

Bu the whole point of my suggestion is using the diff feature to tell what changed, so there is no need to proofread.

Yes we have a nice "compare "feature that could provide a safety net.

Trying it on a French novel, I certainly will get a quick answer for the 10, maybe 20 most used punctuation signs. But punctuation is a world in itself. Who knows how some unusual sign could be rendered ? How some surprising mistakes could later pass unattended?

Is it so difficult to understand that I am afraid that a tool primarily conceived by English-speaking coders (thanks to them) for English books could prove partially inadequate for other languages?

[eschwartz]

Originally Posted by roger64

Yes we have a nice "compare "feature that could provide a safety net.

Trying it on a French novel, I certainly will get a quick answer for the 10, maybe 20 most used punctuation signs. But punctuation is a world in itself. Who knows how some unusual sign could be rendered ? How some surprising mistakes could later pass unattended?

Is it so difficult to understand that I am afraid that a tool primarily conceived by English-speaking coders (thanks to them) for English books could prove partially inadequate for other languages?

It is not a safety net it is a way of picking up every last change in the book. Unicode is unicode, no matter what language you speak or code in.

Use Smarten, see what changed, decide if you like it. Repeat.

Once you have gotten a feel for how Smarten works, you may decide to skip the check, but personally I would eyeball all automatic changes regardless of the situation.

[mrmikel]

Roger64, perhaps you might go to the French forum and recruit some help there. Few of us English speakers could really understand the full nature of what you are talking about. I suspect few of understand the rules that even govern the English part. It should be second nature to the French speaking.

Since calibre is in Python, you might even get someone to help with coding, if that is the issue.

I understand that you, like we English speakers, just want the program to work without having to go through extra motion just accomplish what we wish to do.

[roger64]

Originally Posted by eschwartz

It is not a safety net it is a way of picking up every last change in the book. Unicode is unicode, no matter what language you speak or code in.

Use Smarten, see what changed, decide if you like it. Repeat.

Once you have gotten a feel for how Smarten works, you may decide to skip the check, but personally I would eyeball all automatic changes regardless of the situation.

Yes that was I meant by "safety net", sorry to be unprecise. What surprises me is that this tool is not customizable, and seems to be "Take it or leave it". It may apply a lot of different rules but we have to accept the whole package. This may be good for Unicode, but it seems bad for languages.

Nevertheless, I'll have a try and report later.

[eschwartz]

Perhaps you will be the one to finally add language support to SmartyPants, then.

[roger64]

Originally Posted by eschwartz

Perhaps you will be the one to finally add language support to SmartyPants, then.

Sorry, but I never heard before about these "pants", I only reacted to the word "Punctuation". Maybe if the title had been "smarten your pants", I would have considered the thing differently.

[eschwartz]

Originally Posted by roger64

Sorry, but I never heard before about these "pants", I only reacted to the word "Punctuation". Maybe if the title had been "smarten your pants", I would have considered the thing differently.

Well, regardless of where the smarten functions come from someone's still got to write the rulesets.

You write the French regexes and I'm sure Kovid will add language checking to Smarten.