Firmware 3.19.5761

[Ripplinger]

Originally Posted by drjd

I'm using Calibre v 2.48 to connect the device, but I am getting a "kobo database version unsupported" error. Where exactly should I enable the "Attempt to support newer firmware" option in Calibre, please ?

Preferences > Plugins > Device Interface Plugins > Scroll down and find whichever Kobo plugin you're using: the standard one is KoboTouch, the extended one is KoboTouchExtended. I suggest making the change to both so you don't forget if you ever switch back and forth for epubs/kepubs. Double click the plugin for settings, towards the lower right side check "Attempt to support newer firmware".

[drjd]

Originally Posted by Ripplinger

Preferences > Plugins > Device Interface Plugins > Scroll down and find whichever Kobo plugin you're using: the standard one is KoboTouch, the extended one is KoboTouchExtended. I suggest making the change to both so you don't forget if you ever switch back and forth for epubs/kepubs. Double click the plugin for settings, towards the lower right side check "Attempt to support newer firmware".

Thanks, Ripplinger. Problem solved.

[MacEachaidh]

I voted "no" for satisfaction with the new firmware, but it's for small issues, so it could have gone either way.

For one, the behaviour where a device wakes up a page or so earlier than where it went to sleep appears to be back, at least on my Aura HD. (I haven't tried it on the Touch yet.) Why does this happen, does anyone actually know? Is it specific to certain books? It seemed to have been resolved in the last couple of releases, but now it's back.

And the second is that I'm not thrilled by the direction the interaction with the Kobo store is moving in. I get that Kobo wants to sell books, of course, but I'd prefer access be more optional and less in-my-face. And even then, the functionality of the UI seems to be devolving, if anything.

[emreka]

I see Kobo has implemented Turkish language along with a very small English-Turkish dictionary but there are some translation errors. "Read" (past tense) is translated as "oku" which means "read" in present tense; translation of "%50 read" sounds like "read %50". Also, "sleep/ing" is generally translated as "bekleme/de" (on hold) across platforms such as android, windows, however the kobo translation is "uyuyor" which means "sleeping" but sounds odd in Turkish.

[surquizu]

Originally Posted by emreka

I see Kobo has implemented Turkish language along with a very small English-Turkish dictionary but there are some translation errors. "Read" (past tense) is translated as "oku" which means "read" in present tense; translation of "%50 read" sounds like "read %50". Also, "sleep/ing" is generally translated as "bekleme/de" (on hold) across platforms such as android, windows, however the kobo translation is "uyuyor" which means "sleeping" but sounds odd in Turkish.

You can see in this post how to change translations. If you need help, ask.

https://www.mobileread.com/forums/sho...d.php?t=261771

[BobC]

Originally Posted by murg

Is the new release numbering scheme a psychological ploy by Kobo to make us think that they don't produce so many releases (e.g. are we going to be on release 3.19 for many, many months)?

I think the number is actually the days since 30/12/1999 when work started on the release.

This is what I get using a spreadsheet :
14/05/15 5613
09/10/15 5761
30/12/99 0

The date in column 1 is calculated by displaying the number in column 2 in date format. Day Zero is thus 30/12/1999.

My assumption is that the release number is generated when work on the release starts.

Sorry if someone else has already come up with this explanation.

BobC

[murg]

Originally Posted by BobC

I think the number is actually the days since 30/12/1999 when work started on the release.

This is what I get using a spreadsheet :
14/05/15 5613
09/10/15 5761
30/12/99 0

The date in column 1 is calculated by displaying the number in column 2 in date format. Day Zero is thus 30/12/1999.

My assumption is that the release number is generated when work on the release starts.

Sorry if someone else has already come up with this explanation.

BobC

So ... what you are saying is that Kobo has been working on this particular release for just under 16 years?

[Lynx-lynx]

Originally Posted by BobC

I think the number is actually the days since 30/12/1999 when work started on the release.

This is what I get using a spreadsheet :
14/05/15 5613
09/10/15 5761
30/12/99 0

The date in column 1 is calculated by displaying the number in column 2 in date format. Day Zero is thus 30/12/1999.

My assumption is that the release number is generated when work on the release starts.

Sorry if someone else has already come up with this explanation.

BobC

Wooooot!

[BobC]

Originally Posted by murg

So ... what you are saying is that Kobo has been working on this particular release for just under 16 years?

No - just since 9 Oct 2015 , the previous version would have been started on 14 May 2015.

The dates are formatted dd/mm/YY in the spreadsheet I used - apologies for any confusion caused by date formatting.

The number is the number of days since the reference date when work started so they commenced work on 9 Oct 2015 a date which the spreadsheet (and other systems which count from the beginning of the millenium) expresses as 5761. How long they worked on it would be the actual date of release minus that date.

I use LibreOffice as a spreadsheet - I think Microsoft use a different reference date so others might see different results if they try to reproduce my calculation.

Bear in mind it's just a theory and with only two data points to test on it could well be proved wrong in the future, it does make sense to assign a name when the work starts rather than at some time in the future when it's for for release.

BobC

[davidfor]

Originally Posted by BobC

No - just since 9 Oct 2015 , the previous version would have been started on 14 May 2015.

The dates are formatted dd/mm/YY in the spreadsheet I used - apologies for any confusion caused by date formatting.

The number is the number of days since the reference date when work started so they commenced work on 9 Oct 2015 a date which the spreadsheet (and other systems which count from the beginning of the millenium) expresses as 5761. How long they worked on it would be the actual date of release minus that date.

I use LibreOffice as a spreadsheet - I think Microsoft use a different reference date so others might see different results if they try to reproduce my calculation.

Bear in mind it's just a theory and with only two data points to test on it could well be proved wrong in the future, it does make sense to assign a name when the work starts rather than at some time in the future when it's for for release.

That is far to complicated. The number is simply a build number. It's the number of times they have built the built the firmware since adopting this system.

[PeterT]

It's possible that this is the number of builds from day 1. They will have been using a version control system for a long long time, and all they have done is to include the build number as part of the version number.

[BobC]

As I said it was only a theory based on the two numbers available - I'm happy someone has one that fits better especially as mine was based on a slightly odd baseline date.

It was just coincidence that the two numbers produced dates that were feasible.

BobC

[Lynx-lynx]

Originally Posted by BobC

I think the number is actually the days since 30/12/1999 when work started on the release.

This is what I get using a spreadsheet :
14/05/15 5613
09/10/15 5761
30/12/99 0

The date in column 1 is calculated by displaying the number in column 2 in date format. Day Zero is thus 30/12/1999.

My assumption is that the release number is generated when work on the release starts.

Sorry if someone else has already come up with this explanation.

BobC

Originally Posted by BobC

As I said it was only a theory based on the two numbers available - I'm happy someone has one that fits better especially as mine was based on a slightly odd baseline date.

It was just coincidence that the two numbers produced dates that were feasible.

BobC

Bob, when I read your initial post about this I immediately saw the hilarious side of numbering an update by the number of days that had been worked on it. I appreciate that you are genuine in your thinking, and I do not attempt to cause offence.

[BobC]

Originally Posted by Lynx-lynx

Bob, when I read your initial post about this I immediately saw the hilarious side of numbering an update by the number of days that had been worked on it. I appreciate that you are genuine in your thinking, and I do not attempt to cause offence.

I wasn't offended. It would be amusing to base the number on the length of time worked on it, which actually wasn't what I meant (though as murg seemed to think that was my suggestion my original post wasn't clear enough). I've seen plenty of stuff with numbering like 20140518 referring to the date of a release (May 14, 2014), it wasn't unreasonable to think someone had used a computer equivalent of a date as a means of identifying it.

No doubt someone who comes from the software production world would immediately understand and recognise build numbers; my background is in Project Management (not specifically software related) and usually a project is identified at inception so dates in 2015 seemed feasible in the absence of any other competing opinions.

BobC

[murg]

Originally Posted by BobC

I wasn't offended. It would be amusing to base the number on the length of time worked on it, which actually wasn't what I meant (though as murg seemed to think that was my suggestion my original post wasn't clear enough). I've seen plenty of stuff with numbering like 20140518 referring to the date of a release (May 14, 2014), it wasn't unreasonable to think someone had used a computer equivalent of a date as a means of identifying it.

No doubt someone who comes from the software production world would immediately understand and recognise build numbers; my background is in Project Management (not specifically software related) and usually a project is identified at inception so dates in 2015 seemed feasible in the absence of any other competing opinions.

BobC

Bob,

I'm from the software production world. I understand exactly how programs and operating systems handle dates (usually poorly).

My comment was a joke. Go back and re-read it, especially considering the legendary 'quality' of Kobo's firmware.

I think the date correspondence is accidental. In the software build world, releases are tagged with when they are built, not when they start. I think the latest release started after the finalisation of the Dec release, not in October 2015.

As an example, Microsoft's programming tools use a mangled date to tag a build/release if you use their automatically build numbering system.

The numeric difference between the two releases, 148 for roughly 20 days, suggests that the number is a build number from a CI (continuous integration) system, building approx. 7-8 times a day. Or, for all we know, Kobo only has one programmer working on the firmware, and this is just the incremental build number from the compiles/links.