The most unloved Wimp message... DataSaved

I recently implemented the DataSaved message properly because I clearly don’t have 700 more important things on my to-do list, so of course I did.

It was surprisingly tricky to get right, so I thought I’d check what else bothers… nothing, of course.

Of the standard apps, only Edit/SrcEdit attempts to handle it, and only implements half, which defeats the object somewhat (it’s actually in RISCOS_Lib).

So a quick straw poll, is DataSaved a misunderstood genius or an embarrassing throwback?

1. Very neat feature of the Desktop Save Protocol
2. Utterly useless make-work complexity
3. What’s DataSaved?

Option 3 for me.
So I looked it up – https://www.riscosopen.org/wiki/documentation/show/Message_DataSaved

Yes, that’s hopelessly incomplete (though that’s what RISCOS_Lib implements). OK, so you notice when something else saves your data, but how did the something else know to send you the DataSaved? Undocumented. Where’s the part for noticing where your data came from? Undocumented.

Video

OK, so you notice when something else saves your data, but how did the something else know to send you the DataSaved? Undocumented. Where’s the part for noticing where your data came from? Undocumented.

OK, 701 more important things to do if you count updating the wiki in that particular bit…

Yeah, I know.

I’m going to write a Desktop Save Protocol tester – you’d think there’d be one already. I doubt there’s many ‘correct’ implementations. There’s so many ways it can go wrong.

I don’t think I bothered implementing this in the only application I’ve written which is an editor. It’s not relevant for tools which export stuff in a different format. I think it’s a bit daft, really. I cannot see how it would be useful in normal circumstances. For one thing, the message does not tell your application where the other application saved the file, so it’s not like you can offer the right location if the user edits the file further in your editor and tries to save again.

I wouldn’t bother with it. I wonder whether anyone has found it useful?

I agree about the lack of location. Perhaps defining it as a duplicate of DataSaveAck would have required “too much” metadata to be stored… but it’s a pity it wasn’t. One could retrofit that without affecting existing implementations (if message length>44, new filename is at +44).

I suspect it made more sense it the floppy-only days, as a mechanism for avoiding saving twice when possible. It would make sense for some kind of slow transfer to a remote storage location that wasn’t represented as a filing system, for example.