Skip to main content

Downgrading of the project: Calc has a fatal bug - it cannot reliably link workbooks

The project is to build a Linux Mint machine to have the identical functionality and ergonomics as the existing Windows 10 machine.

This stage relates to on-going work to make LibreOffice Calc workable.

Unfortunately, it turns out that LibreOffice isn't fit for purpose.  Specifically, it cannot reliably link workbooks, even in its own native workbooks.

As a consequence, the project is now downgraded.  My time is now better spent doing other, more productive things, with a greater chance of success.


On 29Jul2018, I complied a detailed test using fresh, sample data created specifically for the purpose, in Excel on Windows 10, then had a virtual machine running LibreCalc on Linux Mint to import it. The basic spreadsheet functionality of linked workbooks failed haphazardly.

I logged the test results on 29Jul2018 on AskLibreOffice and, to date, have yet to see any evidence of anybody taking ownership of recognising the issue and thus fixing it.

The sole response was a false-answer by m.a.riosv, who suggested that it might possibly be related to a bug which appears to be something only vaguely related and substantially less than my findings logged.

Decision: a failure of Calc is, in turn, a failure of Linux Mint

Ultimately, a computer is a complex environment.

It is only as good as its weakest component, and the user determines what that weakest component might be based upon whether the computer delivers the functional requirements.

Being unable to perform the same basic functionality as Excel is a pretty fundamental fault for any spreadsheet package.  Especially when the user's primary app on the computer is a spreadsheet package!

There is no excuse for the basic functionality of linked workbooks to fail.  There are no viable alternatives to linked workbooks.  Especially when there are more than 10 years of Excel data to convert to ODS to be fully functional as-is without the need to manually re-type it just to work in Calc.

But with Calc choosing to perform basic functionality of Excel haphazardly (and badly in other respects!), it means that Calc isn't fit for purpose.

And if Calc isn't fit for purpose, it means that there is a severe limitation in the use of the Linux Mint environment.  Which in turn questions whether the Linux Mint has a viable capacity to replace Windows.

On the face of it, without a viable replacement for Excel, Linux Mint has no viable capacity to replace Windows.  Windows is the only show in town because it hosts the only spreadsheet package - Excel - that is fit for purpose.


The alternatives are:

  • to abandon the project and stay with Microsoft Windows 10, because at least Excel works as required on Windows.
  • to apply a translation layer in Linux Mint, to allow it to run Microsoft Office.
  • re-engineer current and future spreadsheets to avoid all forms of workbook linking.
  • use Google Docs.
On reflection, only the first of these alternatives is viable:
  • to use Linux Mint except for Excel is pointless; the Windows environment is slow, clunky and risky, but at least Excel works in it.
  • a translation layer in Linux compromises:
    • the whole point of abandoning Windows in the first place (might as well stick with Windows);
    • Linux, because the translation layer introduces an attack vector onto a Linux machine that, ordinarily, a Linux environment would not permit.  Linux doesn't really need Windows-style anti-malware software, unless it is compromised either by a translation layer.... or Windows-style anti-malware software.
  • re-engineering current and future spreadsheets would make them less efficient, more risky and result in the effective deletion of an archive of over 10 years.
  • linking workbooks in Google Docs is more reliable than in Calc, but is a wholly different method and thus will be unable to open as-is the archive of over 10 years.


The project is now downgraded.  My time is now better spent doing other, more productive things, with a greater chance of success.

End of post.
Completed 14Aug2018.


Popular posts from this blog

OnlyOffice: keyboard inaccessible, so not useable, therefore not tested

I installed OnlyOffice I had intended to test it with my now-standard test suite of two linked workbooks.

Unfortunately, in spite of a promising look, I quickly discovered that - with one exception - everything was navigable only by mouse.

That makes it a child's toy.  Unfit for purpose!  No point in testing it further.

I uninstalled it within 10 minutes of installing it.

Adjusting screen brightness

The machine on which Linux Mint is installed an old Acer Aspire 5732Z ("Gandalf")

It has buttons to adjust the brightness of the screen's backlight.  When the user uses these buttons, Linux Mint correctly presented a fading-popup box (a slider bar) to denote relative brightness.  But Linux Mint did not actually adjust the brightness of the screen.

It seems to be a known issue in the Linux Mint forums and solved in multiple  stages by the Easy Tips Project.

I followed the instructions on Easy Tips section 5.2 in Gandalf's admin account, then re-booted, then logged in using the user account, and the brightness adjustment function worked correctly.

Easy Tips asks the user to discover the relevant property of the machine, then creates a file that contains a script of parameters that other programs in Linux Mint understand.

This method worked for Gandalf, because Gandalf has an integrated Intel chipset.

Useful commands at the Terminal ALT+T (or the Mint) menu gets to the …

Keepass and KeepassX

The project is to build a Linux Mint machine to have the identical functionality and ergonomics as the existing Windows 10 machine.

This stage relates to password manager, Keepass.
Environment & required functionality A number of encrypted password vaults synchronise between three machines:

The Linux Mint Xfce laptop "Gandalf";The Windows 10 laptop "Legolas";Another Windows 10 machine, name withheld to protect the guilty.
The synchronisation agent is Google Drive in Windows 10, and grive2 in Linux Mint.
Alternatives My original decision to use Keepass was in 2016 and was based on:

Keepass is open-source;Keepass is locally stored, not stored in the cloud;Keepass does not automatically plug into the browser (a plugin permits this if ever necessary);higher security standards at the office, worth deploying at home;portability of the password vault via Google Drive, encrypted such that Google would not be able to slurp data from an otherwise-unencrypted vault.overall …