Offline mode - how to implement it?

When deciding how to do offline mode, there are several trade offs to consider. One way to do offline would be to not allow addition of images or additions/changes to bibliography sources while a user is offline. The advantage would be that it would be easier to implement and that it can therefore ship earlier.

Would such a way to implement offline mode be acceptable?

Hello! For me one important thing is, that you do not loos anything if you loose internet connection, here I think the most important thing is to keep the text, with some placeholders where you might have lost an image or citation.
Next I think, if you consider working offline by intention, you just need to know it beforehand, that some things can only be done online. In writing the text is the most relevant, you always can put the other elements as placeholders. So I would suggest to go without images and citations, so you are able to deliver earlier! Best! andreas

Sure @azitek , under any circumstance you will not lose any text. In case we don’t allow changes to citation sources, the corresponding menu would simply be grayed out so you shouldn’t accidentally start editing bibliography information while offline. You could still insert references to already existing citation sources within the text.

There would also be a small issue when you are in the middle of editing bibliography information when you go offline. I am not sure what to do exactly then - maybe show a warning “You can not save changes to the bibliography source while offline” and gray out the “save” button.

In general it’s a better idea to just be online all the time. I hope Austria and Germany get their mobile networks in rural networks fixed as that really seems to be the main issue for Fidus Writer that are most interested in offline mode.

Would it make sense to implement the offline mode by allowing users to install a local instance of FW with full functionalities, in order to be able to work offline, then syncing it with the online account?

I know this has some drawbacks especially of collaborative editing, but maybe a conflicting versions check could be implemented through track change system or similar.

I am not a programmer, so I don’t know whether this makes sense (please accept my apologies in advance, in case).

@guidogorgoni That’s not really necessary. With Fidus Writer 3.7, you are already “installing” the entire frontend part of the software in your browser - meaning all the JavaScript files, images, etc. are downloaded into your browser and cached there. What is missing in 3.7 is caching application data (the contents of documents, your user name, the bibliography db, etc.) and dealing with merging documents when several users have been working on it simultaneously. The merging document part has landed in the 3.8 development branch, so there is not that much that is left.

Asking users to install all of Fidus Writer will be too big a task for most users. For one thing, it probably will not work unless you run Linux. Also, it requires running a database software, some Java for spell checking and write various things on the command line to configure it. For most normal users none of that is probably an option.

But just to be clear: Fidus Writer is open source and anyone is allowed to install the full software on their own machine. We have made the installation process as simple as possible. But you still need to run Linux and write stuff on the command line and that is most likely too difficult for 99.9% of users.

Thank you very much for the clarification and for the quick answer.

Indeed I am an intermediate Linux user (Kubuntu) and I tried a local install, but for me it needs too much skills for being able to quickly start professional production without taking unreasonable risks.

The solution you are implementing in 3.8 is great, happy to know it!