On Thursday, June 11, 2015 at 5:00:32 AM UTC-5, David Rajchenbach-Teller wrote: > On 11/06/15 03:05, B Galliart wrote:
> > We have done that. It was bug #1172126 [1]. > > Apparently, we do not understand Dan's arguments in the same manner. The > bug you quote is about removing Pocket Integration. As we can see from > the current thread, well, this deserves at the very least an open > discussion, and Bugzilla is an awful medium for that. > > What I read of Dan's arguments is that we should open the feature. This > could be translated, for instance, as the following set of bugs: > - Provide a simple way to erase user data forever from Pocket's servers > [for instance as an item of the "Clear History" button]; > - Give users the ability to use Sync instead of Pocket's servers for > storing their bookmarks [note that this might be tricky, because > providing disk space for hundreds of millions of users is quite expensive]; > - Give users the ability to store/recover their bookmarks using a > variety of protocols (ftp server, dropbox, google drive, ...); > - Publish specifications on the communications between Firefox and > Pocket's servers as an open protocol; > - ... > > Did I misread Dan's comments? If so, apologies. I don't think you misread Dan's comments. That wasn't the point I was trying to make at all. My point is that a product manager of the "user" advocacy group was so quickly dismissive that I doubt if it matter if the initial ticket was about removal or modification. It is was not as if he responded with something like "before discussing removal, lets talk about what can be done differently." The answer was to completely ignore any point being made in the ticket and declare the issue "resolved." If anything can illustrate the degree to which the "welcome to firefox" videos have become empty rhetoric, it would be the handling of that ticket. How is "welcome to personal freedom on the web" at all consistent with a forced opt-in agreement at install to the Pocket(TM) Privacy policy?? If you really think it will do any good to discuss Dan's points, I am happy to do that but I'm not going to bother trying to open a ticket with those points as long as Tyler Downer continues to be a member of the Mozilla Foundation. In terms of the point you brought up: (1) Providing a easy way to erase user data forever (a "Clear History") "Clear History" would not be hard to do with the existing API. However, it may be a lengthy process depending on how many items have been saved. Pocket(TM) requests that the items be pulled in fixed size batches rather than all at once. Also, the API seems to provide all or none, to get the list of item IDs required to delete requires requesting the full item data and not just the item IDs. For the most part, the process of getting the item list for a full clear history would use the same bandwidth as performing a full backup of all the items. The other important thing to keep in mind is the API delete call only promises to remove access to the item from the account. This doesn't claim to do any form of secure wipe on the data from the servers. The Privacy Policy seems to current indicate that the information may get rolled into aggregated data. As such, providing a clear history button may be misleading and imply to a customer that it is providing a greater degree of privacy than actually takes place on the server side. If a feature like this is important, then instead of Pocket(TM), a system which only stores encrypted blobs to the server should probably be used. That way a "clear history" can provide better guarantees to the user by simply throwing away the encryption key and rendering the server side data useless. (2) Give users the ability to use Sync instead of Pocket's servers I think this goes beyond the scope of the Sync 1.5 protocol. It might be nice to discuss how a feature like this could fit into a Sync 2.0 protocol but that seems like a long term goal. (3) Give users the ability to store/recover their bookmarks using a variety of protocols (ftp server, dropbox, google drive, ...) I would prefer this as well except for FTP. FTP over TLS does not always work well over NAT and we should not be encouraging continued use of FTP unencrypted. Not only should the data be encrypted, but the authentication should be as well. I would recommend WebDAV instead of FTP. Given the growing number of free/cheap storage services that have a REST API (even Microsoft OneDrive supports a REST API), it seems that much more damning the built-in method with the browser that brings "freedom" can only work with a single storage provider. (4) Publish specifications on the communications between Firefox and Pocket's servers as an open protocol Pocket(TM) has provided documentation for the majority of the protocol used by Firefox with the exception of the "/firefox/save" call. A contacted Pocket(TM) support for clarification regarding it and got the follow back: "If an endpoint is not on our Developer documentation, it's considered private." This suggests to me that Pocket(TM) is not intending Firefox to use a strictly openly documented API but the use of undocumented calls is by design. That would mean you will have a hard time getting Pocket(TM) to come to the table to discuss a published specification. It also means that future versions of Firefox may use additional undocumented calls which would break any pocket API clones. Unlike Firefox Sync where being able to run a private server was part of the goal, it seem like trying to support Firefox Pocket(TM) will be a moving target of playing catch-up to the API changes each time Firefox puts out an update. How that is consistent with a mission statement of promoting an open web through open standards makes no sense to me. > I can understand why you feel that way. I am personally not very happy > about the fast-tracking involved. But I believe that the best way > forward right now is to take this as an opportunity to: > 1. improve Firefox further; > 2. determine if/where we have made any mistakes and how to not make them > again. > > The possible bugs above are examples on how we could do 1., and a few of > us have proposals in the pipe that may improve 2 (not ready for > prime-time yet, though). This is assuming the Mozilla Foundation is ready to have a two way conversation about this. I think at the very least that Tyler Downer has made it clear through his actions how the user advocacy team feels about having a two way conversation about this. Mozilla Foundation also thumb their nose at the Pocket(TM) Terms of Service requiring agreement at *install* regardless of use by releasing the following statement: "Directly integrating Pocket into the browser was a choice we made to provide this feature to our users in the best way possible. To disable Pocket, you can remove it from your toolbar or menu. If Pocket is removed from the toolbar or menu, then the feature is effectively disabled, though you can still find it again by accessing it in the Customize Panel." While it is true the disabling is an option, the Terms of Service and Privacy Policy both claim to still apply as long as the feature is *installed*. This Mozilla Foundation work-around does nothing to address that. And these actions aren't consistent with the ideals that are promoted in the promotional videos for Firefox. _______________________________________________ governance mailing list governance@lists.mozilla.org https://lists.mozilla.org/listinfo/governance
