Unintended Consequences of Using Save To Disk Feature In Calibre?

[93terp]

In reading a 3 part article about various ways to send books to a Kindle (www.teleread.com), a reader posted a comment about the advantages of using Calibre's Save To Disk feature over using Send To Device. It seemed to make sense to me, and one of the main advantages noted is that using Save To Disk results in less space being taken up on the Kindle. Before I venture down this path (because my Calibre library has 2,100 books & I prefer to have as many of them on my Kindle as possible), are there any known unintended consequences of using this option to load my Kindle? Is there any loss of functionality or formatting associated with this approach?

[dwig]

Assuming that by "my Kindle" you actually mean the K3 you list as one of your devices, there is no downside that I'm aware of. Save to Disk and a manual transfer via USB has been my almost exclusive method of loading my K3.

That said, loading 2100 books on a K3 is generally a very very bad idea. Too many operations on the K3 become painfully lethargic when it has to deal with a library that large. If you must load that many on the K3 should should do it in modest sized batches (10-40 books) and confirm that it has finished indexing one batch before loading a second. Indexing 40 books, particularly large ones, can take many many hours. Don't expect to load more than one modest sized batch per day.

Though I don't use my old K3 that much these days, I now load only a small number of books (~5-10) from my "waiting to be read" list's "read these soon" sub-list and I delete them when read. My "new" device is a Win10 (nee Win8) tablet and I do the same with the reading app that I use, though I have Calibre with my full library on the tablet as well.

[93terp]

Originally Posted by dwig

Assuming that by "my Kindle" you actually mean the K3 you list as one of your devices, there is no downside that I'm aware of. Save to Disk and a manual transfer via USB has been my almost exclusive method of loading my K3.

That said, loading 2100 books on a K3 is generally a very very bad idea. Too many operations on the K3 become painfully lethargic when it has to deal with a library that large. If you must load that many on the K3 should should do it in modest sized batches (10-40 books) and confirm that it has finished indexing one batch before loading a second. Indexing 40 books, particularly large ones, can take many many hours. Don't expect to load more than one modest sized batch per day.

Though I don't use my old K3 that much these days, I now load only a small number of books (~5-10) from my "waiting to be read" list's "read these soon" sub-list and I delete them when read. My "new" device is a Win10 (nee Win8) tablet and I do the same with the reading app that I use, though I have Calibre with my full library on the tablet as well.

Dwig, thanks for the reply. I need to update the list of devices I have, as I'm actually referring to a Kindle Voyage.

[NullNix]

Originally Posted by 93terp

Dwig, thanks for the reply. I need to update the list of devices I have, as I'm actually referring to a Kindle Voyage.

With something that new, in all probability those 'load in batches' rules no longer apply: the indexer has been rewritten and no longer discernibly slows anything down while it's running, and I've never known it get stuck either (though others might have different experiences).

Some operations are slower with really big libraries, but the only really noticeable one other than all-text searches is adding and removing things from collections, which seems to scale superlinearly with the total number of books on the device (crappy algorithm alert!)

[Aqua-Andy]

I have never understood why anyone would want 2000 books on a reader. If I have more than a few dozen loaded I find it to be a pain to find what I want to read. I will usually load a bunch of books that I will read and then delete then when I am finished.

[eschwartz]

Originally Posted by 93terp

In reading a 3 part article about various ways to send books to a Kindle (www.teleread.com), a reader posted a comment about the advantages of using Calibre's Save To Disk feature over using Send To Device. It seemed to make sense to me, and one of the main advantages noted is that using Save To Disk results in less space being taken up on the Kindle.

Says who, and why?

By default, Save-to-Disk performs the same metadata updates performed during Send-to-Device, and other than that, it is the same original file...

Why on earth would there be any difference between the two, except of course that Send-to-Device is device-aware and can select specific formats, use device-specific plugboards, optionally calculate Kindle APNX files, integrate with the Kobo database, etc.

If anything, Save-to-Disk requires more space, because by default it also saves the metadata.opf and cover.jpg (and creates extra folder inodes).

[93terp]

Originally Posted by Aqua-Andy

I have never understood why anyone would want 2000 books on a reader. If I have more than a few dozen loaded I find it to be a pain to find what I want to read. I will usually load a bunch of books that I will read and then delete then when I am finished.

To each his or her own.

[93terp]

Originally Posted by eschwartz

Says who, and why?

By default, Save-to-Disk performs the same metadata updates performed during Send-to-Device, and other than that, it is the same original file...

Why on earth would there be any difference between the two, except of course that Send-to-Device is device-aware and can select specific formats, use device-specific plugboards, optionally calculate Kindle APNX files, integrate with the Kobo database, etc.

If anything, Save-to-Disk requires more space, because by default it also saves the metadata.opf and cover.jpg (and creates extra folder inodes).

So far, I've seen much more free space available on my Voyage after trying the Save to Disk than the Send To Device approach to sending books. By significant margins.

[eschwartz]

Originally Posted by 93terp

So far, I've seen much more free space available on my Voyage after trying the Save to Disk than the Send To Device approach to sending books. By significant margins.

Really? Where did you check that, via the Device settings or the File Browser properties of the mounted Kindle Drive?

Okay, then, go compare the size of the files sent via Send-to-Device and saved with Save-to-Disk.
Because that makes no sense.

In fact, the files should be identical, save potentially for minor timestamp or order-based variation on the exact same data. And if you have set both methods to not update the metadata inside the files, they will in fact be byte-identical.

...

Checking the device settings is tremendously inaccurate, because the Kindle has free rein to do all kinds of processing, indexing, caching, etc. once the device is disconnected, and if all you are looking at is free space left on the device then you cannot tell what is what.
So there are way too many potential causes.

[93terp]

Originally Posted by eschwartz

Really? Where did you check that, via the Device settings or the File Browser properties of the mounted Kindle Drive?

Okay, then, go compare the size of the files sent via Send-to-Device and saved with Save-to-Disk.
Because that makes no sense.

In fact, the files should be identical, save potentially for minor timestamp or order-based variation on the exact same data. And if you have set both methods to not update the metadata inside the files, they will in fact be byte-identical.

...

Checking the device settings is tremendously inaccurate, because the Kindle has free rein to do all kinds of processing, indexing, caching, etc. once the device is disconnected, and if all you are looking at is free space left on the device then you cannot tell what is what.
So there are way too many potential causes.

I've already done both & don't need to do it again. Perhaps you should do it & see for yourself. But more importantly I'm looking for an answer to my original question, rather than a back & forth over whether or not more or less space is consumed. I've already answered that question for myself.

[eschwartz]

Originally Posted by 93terp

I've already done both & don't need to do it again. Perhaps you should do it & see for yourself. But more importantly I'm looking for an answer to my original question, rather than a back & forth over whether or not more or less space is consumed. I've already answered that question for myself.

Not only have I done it, and gotten the results I expected, but the whole logic of calibre is predicated on their being the same.

So there shouldn't be any unintended consequences, for the simple reason that they are functionally identical as regards the ebook files themselves.

But of course you know better. So maybe your unspecified setup which is running counter to my expectations has unintended consequences... how should I know, if you do not wish to participate in determining why the files are different to begin with?

[93terp]

Originally Posted by eschwartz

Not only have I done it, and gotten the results I expected, but the whole logic of calibre is predicated on their being the same.

So there shouldn't be any unintended consequences, for the simple reason that they are functionally identical as regards the ebook files themselves.

But of course you know better. So maybe your unspecified setup which is running counter to my expectations has unintended consequences... how should I know, if you do not wish to participate in determining why the files are different to begin with?

Wow, that's sooo incredibly helpful. Nothing like a know-it-all who provides no help at all, then takes an attitude about it.

[eschwartz]

Pardon me for asking for the information necessary to help you with your problem.

As you have determined I am nothing but an unhelpful know-it-all, I guess there is no point in my sticking around.

> /dev/null

[msel]

Originally Posted by dwig

That said, loading 2100 books on a K3 is generally a very very bad idea. Too many operations on the K3 become painfully lethargic when it has to deal with a library that large. If you must load that many on the K3 should should do it in modest sized batches (10-40 books) and confirm that it has finished indexing one batch before loading a second. Indexing 40 books, particularly large ones, can take many many hours. Don't expect to load more than one modest sized batch per day.

I have also many books (about 2000) on my Kindle 3 Keyboard. In former times I had the following structure of my books on the Kindle:
{author_sort}/{title} - {authors}
My template now is
{series}/{author_sort}/{title} - {authors}
I have found out that the Kindle is more robust when using this document structure. The Kindle need also less time to restart.

[shalym]

Originally Posted by 93terp

I've already done both & don't need to do it again. Perhaps you should do it & see for yourself. But more importantly I'm looking for an answer to my original question, rather than a back & forth over whether or not more or less space is consumed. I've already answered that question for myself.

Hi--are you doing the "save to disk" directly to the Kindle, or are you doing it to the PC first, and then dragging over *only* the book file, and not the cover or the .opf file?

Shari