Drafts functionality - Meta

Jack Douglas replying to PeterVandivier

No it doesn't — and the localstorage 'draft' is distinct from the 'save draft' function, the latter being saved on the database (really just a hidden post until published).

PeterVandivier

i'm curious - does the clicking of "Ask" get the next value from the database sequence and then it's just cached in localstorage with the draft? i'm picturing fun Josh might have in dev tools getting `@@@ question 2147483647` or something :laughing:

PeterVandivier

![blob](/image?hash=4649b7c099b2c2674f5a6291ab9523cf5fddbc66d5586bf73f5dddd6f1e46e1a)  
  
☝️ this is  of fun :-P

PeterVandivier

@@@ question 1496

PeterVandivier

has anyone noted what happens to the chat of a draft on-publish yet?

i-one replying to Jack Douglas

I had 60 revisions this time and was curious to see how confirmation look. Ok, no problem, next time may be.

Jack Douglas

sorry for the delay

Jack Douglas replying to i-one

correct, it's still on my todo list!

i-one replying to Jack Douglas

@Jack, it looks that "Purge Drafts" still has no confirmation.

Jack Douglas replying to i-one

Good point we should disable that you are right

i-one

@Jack, the other odd thing is chat. I think it should be hidden or disabled for the drafts. I can't imagine (mentally health) person that would want to have self discussion of own draft.

Jack Douglas

@i-one FYI, drafts are now excluded from the recent question list and from searches. You can find them on your profile if you need to though.

Jack Douglas replying to i-one

both excellent points :)

i-one

@Jack, two additional observations regarding current drafts workflow: 1) it seems that content license can only be selected at first draft save, it should be possible to change it anytime before publish I think 2) it seems that there is no way to delete draft that was saved, if I change mind to publish

Jack Douglas replying to i-one

Thanks for the prompt — I forgot to add a confirm dialog with a big warning that drafts are lost forever, I will add that at some point soon…

i-one

It appeared to me that there was no confirmation yet, right?

i-one replying to Jack Douglas

Seems working. I clicked "Purge Drafts" button and drafts evaporated. Who knows what happen to them though, may be they fall to Jack's basement :)

Jack Douglas replying to i-one

It is now available :)

i-one replying to Jack Douglas

Also, it would be strange probably to hide something presumably valuable.

i-one replying to Jack Douglas

My drafts are just garbage after publishing, and can be deleted with no regrets. It's like joinery works, who need all those sawdust and wood shavings after job is done? I'm accustomed to cleanup after myself, so I would prefer cleanup button. There should be confirmation message probably to prevent accidental deletion for those who feel that drafts are valuable.

Jack Douglas replying to i-one

I just started work on this and had a thought. Rather than deleting draft revisions, how would you feel about simply not displaying them by default? Then we just need a 'show draft revisions' feature for the rare case when they are actually needed after publishing.

i-one

I'm eager to be the first person to test _Cleanup draft revisions_ when it will be available.

i-one replying to Jack Douglas

Thanks, Jack!

Jack Douglas replying to i-one

you are right of course — now fixed!

i-one

Shouldn't times be related to publish moment?

i-one

![blob](/image?hash=4db3d2dc2e0399fcf4abe0d7e5d380a0b677bc36e618b29ae7ffebea000420db)

i-one

@Jack, tried Drafts with question today

i-one replying to Jack Douglas

Yes, it seems the right place.

Jack Douglas replying to i-one

I was also thinking that the latter might be less intrusive — and I guess the obvious place to add it is on the post history page?

i-one

@Jack , I can't imagine a reason to keep drafts history after publish so far, but may be someone else will find it useful, who knows. Can we have a checkbox "Keep drafts history" (on by default) so that one could uncheck it and clear draft revisions on publish? Or would it be possible to cleanup draft revisions _after_ publish? I'm thinking towards the latter.

i-one replying to Jack Douglas

Thanks, Jack. I'm going to test drafts functionality this month probably.

Jack Douglas

I'll say more when I post an answer in a bit…

Jack Douglas

Plus I've been working on it all day and am just about to release!

Jack Douglas replying to i-one

I agree, but I've been thinking about this a long time — and I'm reasonably confident (though not 100% — that only can come with post-implementation feedback!) that it is the correct way to go.

i-one

@Jack, my experience is telling me that good architecture often reflects nature of things. It is good, in general, if existing code base can be reused. But don't be seduced to abuse it, if it doesn't fit exactly, because of eventually you may realize that model dictates reality instead of serving to it. It may lead to undesired coupling and unjustified restrictions in future. Given that temporary solutions tend to become permanent also, I suggest you to put it aside for a while (as there is no rush on this) to mature and to weight implications of possible alternatives. I know it from my own experince that implementation goes smoothly, when critical mass of understanding the problem is reached.

Jack Douglas

As the solution I'm proposing is so easy to implement I think perhaps we should try it out — we have the option of tweaking it or doing something more complex later if it proves not good enough…

Jack Douglas replying to i-one

That sounds feasible enough, but it doesn't leverage any code we already have in place, and would mean another thing to maintain. Local storage also doesn't always work: in private windows on iOS for example. As you also say in your post, it's tied to each device/browser you are using.

i-one replying to Jack Douglas

The warning could be such that instead of clicking Yes button you have to type "Yes", and then you have to type "Really" second time :D As to the harm, who knows, increasing CO~2~ emission and heating the planet may be.

i-one replying to Jack Douglas

Local storage is the black sheep here. What if draft would save to some "current" at the top of history stack every minute for example? Or probably local storage could serve as a buffer, and some timer could post draft to "current".

Jack Douglas replying to i-one

Would it do any harm to keep them? I imagine that if we delete them on publish someone will click through the warning and then miss them. Storage space isn't an issue.

i-one

Regarding #2, I don't feel the need (yet) to keep revisions prior to publish. I'm Ok with removing draft and all snapshots on publish. But probably it would be nice to warn about it.

Jack Douglas replying to i-one

Every time you manually hit the 'update draft' button (which will be the button text instead of 'update post' when the post is unpublished). You would only need to hit that button when you want to save a snapshot — there would still be the local storage mechanism for unsaved changes as there currently is (with all it's limitations)

i-one replying to Jack Douglas

How draft history is supposed to be formed? Is it that historical state is added every time some "Save" button is pressed? In this case I will probably generate hundreds of historical states per post, which is overkill probably. The benefit of manual snapshots is (presumably) space saving. I feel that it is unlikely that I will ever need more than ten snapshots per post.

Jack Douglas

The benefit (to devs) of this kind of approach is that implementation is almost trivial

Jack Douglas

@i-one, re your update, how would you feel about:  
  
1. Full draft history is retained in exactly the same way normal edit history is retained  
2. Revisions prior to 'publish' are only visible to post author  
  
This would mean snapshots are not automatic — only occuring when manually triggered, would that work do you think?

i-one replying to Jack Douglas

To be sure we understand _'edits'_ in the same way. I imagine it as you open draft (created earlier), change it and save and can return to it later. Is it what you meant? In this case yes, I feel the need to have it.

i-one

Either no history, or purged on publishing I think. Personally I don't feel the need to have history for drafts. Draft is not full-fledged post yet, it is some staging, imo. So it may not need to have all attributes inherent to posts. It may propagate implementation-wise as well, e.g. drafts may even live in a separate table (to not poison existing things).

Jack Douglas

also, do you think that this should just be for new posts, or should draft 'edits' be allowed? The latter would make this more complex to implement

Jack Douglas

@i-one how do you see post history working — should all 'draft' revisions have no post history, or should history be purged when publishing, or just always only visible to the author, or visible to everyone?

1 Answer