Issues Arising - Backup to help UNdo

[@TiddlyTweeter]

It's interesting working on Polly, A PowerShell framework for helping manage TW5 & TWC files.

Some issues came up that in TW I don't think I would have thought about before. 

But I am now.

Here is one of them ...

TiddlyWiki does not have an "undo" function.

 

This could be partly addressed if backup plugins could ALSO restore.

Why? Because "restoration" is the only "undo" we have.

IF it could be better automated it would be very practically useful, I think.

TT 

 

 

[@TiddlyTweeter]

Repeat for email ...

[TonyM]

Josiah,

I think my experience now with tiddlywiki is approaching 10 years. Starting with twc and prior to the tiddlywiki apocalypse, which never came like most apocalypse dont, I had a few saving or corrupt changes disappointments but the community response to the browser challenges have delivered far more than I can imagine. My own practices now garentee recovery of anything I want, at a tiddler level, bunch of tiddlers or whole wiki level.

Tw reciver, timimi and tiddlyserver all have reliable backups to mention a few. I am not so sure on folder based wikis. Maintaining automatic save performs well.

I think what you are suggesting is here but a plugin and better ui is all we need. One proposal I think would help is to simplify filtering tiddlers from the import mechanisium.

There are a few problems that can only be fixed with a direct edit of the wiki file, but good practices can protect you from that.

What do you perceive to be the gap?

regards
Tony

[@TiddlyTweeter]

Ciao Tony

My query is about levels of automation.

I fundamentally agree that PMario's and Riz's plugin systems are excellent.

I simply wondering about RESTORE PROCESS. Whether they can be automated in some way to not just save, but restore? 

It is a valid thought.

The originating observation was that TW has NO native undo, though Mohammad's recent Trash Bin is a real step towards less worry.

I am not expecting anyone to solve the issue. Merely noting it exists and its a real limit in TW.

TT 

[TonyM]

Josiah

I understand your desire for simplification. I think a good design approach would be if the backup process kept track of the time and date and filename of the backup or allowed you to browse them then provided a UI for selective recovery.

Actualy a smarter idea may be a special edition wiki designed to import as json bundles, any wiki and one or more of its backups where you can sort filter inspect review generations and differences and choose what you want to import. Perhaps even letting you merge two or more tiddler versions to create the new version you drop on your primary wiki.

You could call this a wiki tiddler generation management tool. Such an abstracted tool could easily be used to build additional solutions such as undertaking a historical view of a wikis changes, using it to study changes in transactional data and a few more ideas in the back of my brain.

I see no reason we need to keep putting more into our wikis when good universal tools can be held out side and achieve more. After all it is very easy to transfer content between wikis and getting better all the time.

By the way leveraging local storage to retain something you may want to undo in the current session and also allow you to export current session data as a kind of overlay backup also has some merit.

Regards
Tony

[coda coder]

On Saturday, August 17, 2019 at 8:49:55 PM UTC-5, TonyM wrote: 

By the way leveraging local storage to retain something you may want to undo in the current session [...] also has some merit.

Now that's inspired. Thanks Tony.

 

[@TiddlyTweeter]

TonyM: What do you perceive to be the gap?

 

Simply, lack of a clear restore process.

TT

[@TiddlyTweeter]

TonyM: By the way leveraging local storage to retain something you may want to undo in the current session 

I assume by "Local Storage" you mean IN-browser storage? Right?

I myself think its useful in discrete conditions, but I don't want to trust ONE browser. Better to have TW that exist as, on-save, as one, always integrated whole.

By "discrete conditions" I mean specific purpose of short duration. Like a user being able to hold a half-completed questionnaire until done. 

Its VERY good in the sense that an occasional user could save without ANY infra-structure but a browser. That has seriously NOT been talked about. The user who needs top do ONCE. Finish.

I spent a lot of time with NoteSelf, which is "Local (in browser) Storage" PLUS remote database sync. Excellent approach, and robust.

But like so much on TW, if you have few users it gets very hard to develop anything further.

What is the user base for such innovations as NoteSelf & TonyM?

TT