What if only Firefox had write access to that temp folder. So, if the user made changes to the document, and hit ’save’, Word would be forced to ask the user where to save the file.

Yeah, I think that’s actually a good idea. The file could be marked as read-only by Firefox, which should solve the problem.

Another possible way to avoid hassling the user with a click-thru dialogue each time would be to have an option for firefox to remember the user’s preference for handling certain file types

I think this is the case already, but I’m not sure that’s the best idea, because then you have two different kinds of behaviours happening for non-obvious reasons (the user might not remember the decision that they made).

An example: http://thecongo.co.uk/picture2.jpg

This way, the user gets a degree of control over where the file is saved, and won’t get prompted again in the future.

(Apologies for the awful mock up)

You click a link for, say, a Word doc.

Let’s assume for a second the user wants to edit it right away, so it gets downloaded to a temp folder and opens.

What if only Firefox had write access to that temp folder. So, if the user made changes to the document, and hit ’save’, Word would be forced to ask the user where to save the file.

I’m not even sure if this could be done, but it seems that if we are going to resolve this issue without requiring any kind of click-thru when downloading files of this type, then we have to force the external app, i.e. Word, to do some of the leg-work and bother the user with a save dialogue.

Thoughts?

Your mockup is interesting — it gets rid of the hidden consequences of the open vs. save choice, and makes the choice a bit easier. I still think it would be nice if you didn’t have to click through a dialog every time you download a file though. Your solution also assumes that it’s okay for all files to be downloaded to the download folder (often the desktop)…but maybe that’s not a bad thing.

Here’s a mock-up of what I was getting at:

http://thecongo.co.uk/Picture1.jpg

Although perhaps the emphases should be placed on ‘Download & Open’

|_________ File Download____________| | | | You clicked on a link that | | downloads a file. What would | | you like to do? | | | | [Download & Open] [Just download]| |____________________________________|

Yeah, I know what you mean. Certain kinds of files are (likely) ephemeral, and those shouldn’t be treated the same way as a PDF or Microsoft Word document. But it’s kind of a hack to explicitly specify which category each file type falls into, especially since you could almost always come up with some kind of exception.

I still think that defaulting to “open” is the best way to do it. Then the semantics of following a link are always the same: the linked resource is opened (possibly in an external viewer). I do agree with you that there are certain situations where you don’t specifically want to save the file and NOT open it, but I think that those are in the minority. And for those cases, there is a different action: right-click and “Save Link As”. Anyways, having a file open shouldn’t prevent scp from reading it, should it? Although I know that Window isn’t nearly as sensible as Unix about some of those things (e.g. deleting a file that is still open).

Then I started thinking about the stuff that I download. A .torrent file isn’t the same as an OpenOffice document. I don’t want .torrent turds clutting up my download folder, but I don’t want to have the same moment of terror your roommate experienced. Ephemeral file types shouldn’t be treated the same way as long lived files. So we have to ask how a filetype is declared as ephemeral. I’d be happy to add a bit to the file associations in the browser.

Tangent: I take issue with your comment “If you’ve clicked on it, don’t you probably want to open it?” I occasionally buy stock photography from iStockPhoto. When I download the zip file the first thing I want to do is copy it to offsite storage. It cost me money, so I delay my gratification for a few moments to do the scp. Since I’m using Windows, I don’t want any other processes touching the archive, because they might prevent my SSH client from reading the file.