Removing Unused ID=" " 's?

[phossler]

Is there a clever RegEx way to remove any unused (aka 'orphaned') ID's?

Possibly [Check Book] could handle it with a warning that ID="filepos991221" is not used. Remove or Remove all unused ID's??"

Sometimes I want to cleanup an epub and it seems like almost every <div>, <a> and etc. has an ID on it

I know they probably don't hurt anything, but 1) I like to be tidy, and 2) they do make the file a little larger

[kovidgoyal]

ids are not just link targets, they can also be used by CSS. They are also often used by bookmarking/annotation tools to make bookmark locations more reliable. So generally speaking, you should not be removing them.

[JSWolf]

I remove unused IDs and I've never had a problem. But you do have to be careful and not remove any IDs you think are unused to find out there are links to some of those IDs.

[phossler]

@kovid - I understand the risks, but I just got through cleaning an epub that was littered with ID's that seemed to be paper book page number markers

@JSWolf -- if I'm not sure, I Cut it and then run [Check Book]. If I'm wrong, I can Paste it back.

Still it's tedious, and the computer is SO much more accurate than I (hint, hint)

[kovidgoyal]

@phossler: It is for things like this that I made a plugin framework for the editor. There are many things that I dont think are important enough to spend time adding to the core.

[phossler]

Understand -- Not a problem

There is some posts I was reading about a Merge ePub PI being enhanced to include cleanup things like unneeded <span>....</span> pairs, etc.

Will your eventual S&R using functions (I believe that's what you called it) be able to find "ID=".*" and return the Count and if Count=1, replace it with 'Nothing' ?

[KevinH]

Originally Posted by phossler

@kovid - I understand the risks, but I just got through cleaning an epub that was littered with ID's that seemed to be paper book page number markers

Instead of removing the start of printed page ids, you could instead build a page-map.xml from them or use the in the ncx with pagelist tags and add back the original support for proper pages. This would be an easy editor plugin.

Alternatively before removing them, you should probably look for a pagemap.xml file and examine the ncx file to make sure they are not referenced there, otherwise you will break the page mapping feature.

KevinH

[JSWolf]

I don't find page mapping to be all that useful. ADE's page numbering works well enough to have an idea of where I am in the book, how large the book is, and how much more I have to go.

[KevinH]

You have obviously never had to universally cite a reference to a particular passage in the book that people reading on other devices or in print can actually find.

This is common in academic settings and one of the reasons for supporting print page mapping in ebooks.

Originally Posted by JSWolf

I don't find page mapping to be all that useful. ADE's page numbering works well enough to have an idea of where I am in the book, how large the book is, and how much more I have to go.

[JSWolf]

Originally Posted by KevinH

You have obviously never had to universally cite a reference to a particular passage in the book that people reading on other devices or in print can actually find.

This is common in academic settings and one of the reasons for supporting print page mapping in ebooks.

I don't read reference books on an eInk reader.

[PeterT]

Originally Posted by JSWolf

I don't read reference books on an eInk reader.

+1

In fact I would think that even if I *DID* I would still get a hard-copy edition and provide the citation using it's publisher data as opposed to that of the eBook edition.

[KevinH]

Originally Posted by JSWolf

I don't read reference books on an eInk reader.

Wow ... and that's relevant how?

FYI, citing passages by page and edition is not only for reference books.

Scholars often study the works and letters of famous authors in fiction and non-fiction from multiple time periods and need to cite passages.

And more generally, anyone simply may want to cite events or dialog from any book the are reading in parallel with friends many of them who read the print edition.

[JSWolf]

Originally Posted by KevinH

Wow ... and that's relevant how?

FYI, citing passages by page and edition is not only for reference books.

Scholars often study the works and letters of famous authors in fiction and non-fiction from multiple time periods and need to cite passages.

And more generally, anyone simply may want to cite events or dialog from any book the are reading in parallel with friends many of them who read the print edition.

If there are multiple print editions with different age numbers, how will you know which print edition the page map uses? You don't as it doesn't say. Also, most eBook do not have a page map so for ePub, the only page numbering you have is that of ADE. So really, if you are going to be citing from multiple books and you feel you need to have a page number that aligns with some print edition, the forget eBooks.

[eschwartz]

Originally Posted by JSWolf

Also, most eBook do not have a page map so for ePub, the only page numbering you have is that of ADE. So really, if you are going to be citing from multiple books and you feel you need to have a page number that aligns with some print edition, the forget eBooks.

Wait... so you agree with him?

[KevinH]

It's called metadata for a reason. Simply fill in dc:source with the isbn number of the specific print edition.

And, yes real commercial ebooks do this even for popular press fiction. for example most of the last 5 or so Jim Butcher ebooks do this quite nicely.

Originally Posted by JSWolf

If there are multiple print editions with different age numbers, how will you know which print edition the page map uses? You don't as it doesn't say. Also, most eBook do not have a page map so for ePub, the only page numbering you have is that of ADE. So really, if you are going to be citing from multiple books and you feel you need to have a page number that aligns with some print edition, the forget eBooks.