5 ms·

I've stopped using mine, removed all apps, unfollowed (er, unliked?) all pages. I'd guess the market is ripe for a fb-type site that, for lack of a better phra

I've stopped using mine, removed all apps, unfollowed (er, unliked?) all pages. I'd guess the market is ripe for a fb-type site that, for lack of a better phrase, isn't evil. And doesn't have a horrible ux.

pg 16y ago
What features should it have? (Or not have?)
- philcrissman 16y ago
  I mentioned a few things that bug me in the above reply-to-myself... Should: - Be able to export all friend/contact information anytime. CSV is fine. - Be able to _easily_ block/remove applications & notifications. It's fine people can create "which star trek appliance are you?" quizzes, but let's make it _easy_ for these to be removed from my feed if I don't want to see them. - Maybe have a more twitter style follow where it does not need to be a mutual follow. Obviously less information is shared if I don't follow back; etc. Just pondering this, it might or might not be better. I actually don't think there's _that_ many things. There are a few key things that, if done right, would make it a better facebook. I'm just stabbing in the dark though. Can this post be considered a YC application? ;-)
  - TrevorBramble 16y ago
    Better than removing the deluge of unwanted app notifications individually would be a list, off to the side, of apps your friends have used that you haven't opted in or out of yet. Opt in and they all show up, opt out and it's added to a list of unwanted apps in your profile that you can edit later if desired. Obviously this would make users happy but not app makers, because the way it currently works is free (obnoxious) advertising for them. The mixed-access to updates based on bi- or unidirectional association is one of the very nice things about LiveJournal's social scheme. I knew a few people who use FaceBook for one level of communication and retain their LiveJournal accounts for another, because of this.
- ihodes 16y ago
  Those basic ones allowing communication: essentially directed and non-directed asynchronous sharing. And those you 1) opt-in for 2) pay for. There's a long list I have, but that's a startup for another day. (Or for a time a month from now... when school lets out.)
- davi 16y ago
  1: I want ACLs meet Web. I want to publish personal stuff to close friends & family, professional stuff to colleagues, and random miscellany to the world, all via the same tool. I want to be able to customize the layout of my content depending on the audience. 2: I want this without my friends and family having to learn about OpenID or having to sign up for a new account somewhere. (I don't know how to do this, I am just saying what I want. Posterous got this part right.) 3: I want to know that the investment I'm making in generating content will last. This means I need to trust that the host of my content is not going to disappear, or get bought and change its behavior radically. Alternatively, it means that I am given the ability to migrate my content to a well-behaved, API-implementing host elsewhere (including, possibly, on my own server). Nothing I know of addresses this set of wants elegantly. Maybe because monetizing the social graph requires as much transparency from the participants as possible, at least according to current thinking. But I feel that eventually, what 'sharing on the web' means will inevitably evolve to accommodate this set of wants. If Facebook decided to be less evil they could do ACLs-meet-web pretty easily, since they already have such a huge installed user base. In fact, they may have already done this a long time ago for all I know; back when I canceled my account, they did not have this capability.
  - gcr 16y ago
    You are describing the Internet. Self-run web sites, blogs, and instant messaging systems: 1. allow you to fine-tune who sees what, 2. do not require extra accounts (save comment systems, etc.), 3. are not tied to any one provider or company and thus will last for a long time.
    - davi 16y ago
      How do you see the internet providing built-in ACLs? http://en.wikipedia.org/wiki/Access_control_list http://en.wikipedia.org/wiki/Access_control_list Even if it did, your assertion is like equating assembly to a high level language. Just because you theoretically can do anything in assembly that you can in a high level language doesn't mean you would want to. I am trying to specify a set of attributes for a non-existent tool that would let me more fluidly publish what I write and create. ---- edit: if anybody reading this would care to explain why I am getting downvoted & the parent is getting upvoted so much, I would appreciate it.
- codexon 16y ago
  Another interesting question would be: how will it make money? Websites like Facebook and Myspace are shunning the simple advertisement model. Selling private data will become the new standard. A website similar to scope and feature like Facebook will probably not be able to sustain a positive cash flow or break even without selling people's data. UNLESS...they cut features like photo storage, or make a technological breakthrough. They are making ~$2/user/year (not including expenses). Since it recently supposedly became cash-flow positive, it would be reasonable to assume that the expenses are also around ~$2/user. The question then becomes, how much of this $2 is slashed off when you decide to not sell private data? You only have to look at the cost of conventional CPM ads (usually less than $0.01 per impression) and multiply it by how many times a Facebook user opens Facebook each year. My bet is that it is far less than 365/2 since Facebook claims 50% of "active" users login every day.
  - nfnaaron 16y ago
    "Websites like Facebook and Myspace are shunning the simple advertisement model." ? Every one of my facebook pages (feed and profile) always has two or three prominent ads over on the right.
    - codexon 16y ago
      What I mean is they are selling your personal information __in addition__ to regular ads (which are selected by datamining your personal information). As opposed to just regular ads.
  - kmavm 16y ago
    Can you provide an example of facebook "selling private data"? I work there, and it would be a great surprise to me if you could.
    - codexon 16y ago
      Sure. http://en.wikipedia.org/wiki/Facebook#Beacon http://en.wikipedia.org/wiki/Facebook#Beacon The prime example that resulted in a lawsuit settlement. I mean really? Opt-in to reveal your purchasing and browsing habits with selected companies and then spam your friends? http://gawker.com/5426176/facebooks-great-betrayal http://gawker.com/5426176/facebooks-great-betrayal Another opt-in that is equivalent to Facebook buying web coverage at the expense of user privacy. http://gigaom.com/2010/04/22/facebooks-instant-personalization-is-the-real-privacy-hairball/ http://gigaom.com/2010/04/22/facebooks-instant-personalizati... These websites are able to scrape your data to provide you a "personal" experience. Another opt-in. I'm not sure if FB is selling this api, but I'm willing to bet its a certainty. These might not literally be bundling up personal information and selling it outright, but given the track record here, I wouldn't be surprised if they did another "opt-in".
- nfnaaron 16y ago
  I think Google could do this entirely with blogger, maybe reader, maybe Picassa, and email, and a bit of code. It would be a huge boon to Android phones. Google groups is broken for some people, and I'd stay away from using it as a building block. Lots of people have negative opinions about Wave, me included. I don't even know if Buzz would help in any of this, the launch was so botched I never looked at it; people have told me it wouldn't work well for anything like this. Posterous or Tumblr could do it too. Google doesn't seem to be very motivated in integrating their various acquired apps; maybe this would be that motivation. Or maybe Posterous or Tumblr might find some motivation in this. It need not, maybe should not, and probably could not compete with Facebook. That's not a reason to not do it. Jones Soda does pretty well, yet I doubt if they think of themselves as a Coca Cola killer. Whether Google or someone else does it, I'd like: GENERAL - Total control. I don't want to be a control freak, I want to be a control super freak. I decide exactly what is private, shared and public. Everything that can be shared in any way should be obvious, and the settings interface for controlling privacy should be very obvious in: a) what setting controls what feature, b) what feature is controlled by what setting. If you have to guess how to change a feature's exposure, or what feature a setting controls, then you don't have control. - Long-lived settings. New features should never expose tomorrow what was not exposed yesterday. That's just so disrespectful on facebook's part. - New features should begin with the most restrictive privacy settings. - No apps or third parties. Make your money through advertising or paid features. If you're Google then you already have great advertising revenue and this just ties into it. - No friend can expose anything of mine to any greater degree than I have explicitly exposed it. DISCOVERABILITY - If I want to be invisible at any level, then I should be invisible as long as the rivers run and the winds blow, until I change it. Invisibility includes: - The outside world. I'm either discoverable to anyone outside the system, or I'm invisible, according to my settings. - People inside the system. I'm either discoverable inside the system, or I'm invisible. - My groups of friends. I'm either discoverable to my friends' groups, or I'm invisible. EXPOSURE - I control what parts of my data are exposed, and how far out. This is both independent of discoverability and cooperative with it. For example, no matter how far out I might expose a photo album, it can't be exposed farther out than my discoverability. On the other hand no object needs to be as exposed as my maximum discoverability. If I'm invisibile outside the system, then I can't set the exposure of a photo album to outside the system. If I need a photo album or something else to be exposed farther out than my discoverability, then I'd need to create that album outside the system (say, as a picassa album in a different, public account). I might be able to link to such an album from inside (and that link would have the same discoverability and exposure protections as any other object). FUNCTIONS - Discoverability. Anyone can discover me, as long as they're within one of my discoverability spheres. If I've set my discoverability to "world" then you can find me through search or browse. You can send a message to me if I've enabled that (with enough hurdles to discourage spam). If you're a member of the system you can request to friend me. If my discoverability is more restrictive, then you might only be able to discover me if you're already a member. Or you might be able to discover me only if you're a member of one of my friends' other groups. - Exposure. Objects that you post or share have a settable exposure level. Posting objects to a group gives those objects the exposure that you previously associated with that group. - Chat. - Alerts. If a friend logs in while I'm in, I want to know it, unless they've set that data to hidden. - The usual. Can post comments, pictures, music, books, links, "other stuff" and have conversations about them. Post things about yourself in your profile, with different parts of you profile's exposure individually controllable. Be alerted to comments in conversations related to you or your posts, and changes to posted objects. - A group of one. I should be my own group, similar to how some Linux distros create a group whose name is your login and whose membership is just you. I may want to keep track of things just for myself. I may want to stage new additions, initially to just me, and then exposed to a gradually wider audience as I see fit. I may want to pull some things in to private that were exposed. Everyone starts as a member of their own group of one ("self" ?), unexposed and undiscoverable. You have to explicitly make yourself more discoverable, and you have to explicitly expose any objects in your group of one. - Explicit groups of two or more. I should be able to explicitly band together with other people and form a group. Everyone in the group is aware of everyone else in the group. We can all post, converse and share. The group can be invisible or discoverable. I can expose my membership in the group to any group outsider who can see the group, or I can hide my membership from anyone outside the group. The email analogy to this might be cc:. - Implicit groups of two or more. I should be able, on my own, to place people and groups inside of implicit groups that only I am aware of. If I post to an implicit group, each individual or group member of an implicit group receives the post or other object, and can carry on a conversation or other activity related to the object, but no individual or group within one of my implicit groups is aware of any other members of the implicit group; they aren't even aware that they themselves are part of one of my implicit groups, although they might guess from my style of writing. The email analogy to this might be bcc:. The names of the two previous group types, explicit and implicit, may not be the best names. I initially described them as public and private, but those terms had too much confusing overlap with the notions of discoverability and exposure. - Some groups owned by individuals and some groups owned equally by members. Not sure about this. Worth thinking about. - Group union and intersection. I'm a little fuzzy on this one. Say you broadcast post an object to two groups. Person A is a member of one group, B is a member of the other group, yet the two people A and B have no relationship to each other within the system. If people A and B's discoverability settings would allow it, then they should see each other's comments related to the posted object and be able to take part in that object's conversations with each other as if they were both members of the same group. If their discoverability settings would not allow either to discover the other, then they should not see the others' comments. If this function were done poorly, then gaps in conversations related to an object might be disorienting. - Acknowledgement of the outside world. Under my control and choice, I should be able to make automatic tweets, or posts to my external blog, etc, based on what I post and where inside the system. I should also be able to make incoming posts from the outside, like ping.fm to twitter and facebook. - Data migration. I should be able to export part or all of my data quickly and easily. - Data ownership. I should be able to change the exposure of data that I post, and to delete it. - Existence ownership. If I leave, I should be able to totally delete myself. Or, if I want, I could de-activate myself. But it shouldn't take searching outside the system to find a little known method for deletion. If I delete posts or data that are part of conversations, the system should insert a placeholder, but if it's mine I get to kill it. IMPLEMENTATION For this part I'll assume Google is implementing it, because they already have all the parts and we're all familiar with them. Applies equally to Posterous, Tumblr or a player to be named later, acknowledging that they would have more to build. The fundamental activity in the system is posting objects and having conversations about them. An object might be text, media, a link, etc. Posting objects could be handled very well by blogger. You can put pictures and links in a post, and there's a comment mechanism. The comments would continue the conversation. You could also handle posts and reading with gmail, because it has excellent conversation threading. If you use gmail you'll have to be very careful about leakage, in or out. Slashdot's forum has excellent threading, for inspiration. Alerts and keeping up to date could be handled by logging in a lot and reading your stream, or via Google Reader, or by email/gmail, or any combination of them. Again, be very careful about leakage in and out. Be nice to be able to post via email/gmail (hi Posterous!). Media and albums by Picassa, but not just photos, anything a member has the right to post: books, audio, video. Implementing some of this might be a challenge. Pick a subset. Iterate. UI, UX, UXB If you find anything looking like any part of Facebook, you're probably doing it wrong. User interface should tend toward minimal. There should be zero or one way to do everything. UI elements should not suddenly disappear, to be found again after a frustrating search. Facebook recently made the logout link hard to find, burying it in the Account menu. Probably because they want to discourage you from logging out. Don't do stuff like that, it's disrespectful, and even if there isn't evil intent it encourages discussions of evil. You should be able to easily choose where anything will be posted, and not have to worry at all that you might not understand where it's going. When you view anything, you should get some indication of who else can see it, general by default but able to drill down to groups and individuals if you like. Name everything well, once, and use only that name, everywhere. You shouldn't need a help system. But someone will want it, and it should be worth reading. You should also be able to get to help for specific functions, from the context of that function, without having to abandon your context and go into "reading the manual" mode. There's probably more.
- tmsh 16y ago
  They shouldn't be so successful. They made the same mistake as Goldman. In this case, small, problematic privacy features (people being able to tag you without your permission, etc.) that weren't exhausted + ambitious new platforms and exponential success suddenly caused them to reach a critical mass where people think twice about their service. It's all about the integrity of the core feature set, and avoiding future, short-term successes that erode into it. Ironically, one only really notices the erosions (which, as with Apple and Goldman, are really not unique within their industry) after a certain amount of success.
- alecco 16y ago
  1 Reasonable privacy 2 Plausible deniability (like OTR) 3 Fast updates 4 Easily customizable 5 Client-side extensible interface (Greasemonkey style) About reasonable privacy, the service shouldn't be able to read content. The provider can tell Alice (user on IP 1) sent something to Bob (user on IP 2) but the message content itself shouldn't be compromised. In particular profile-identifying details should be reasonably safe. This requires client-side encryption, of course. Plausible deniability isn't that tricky, it mostly requires careful use of protocols at design phase and there is OTR to borrow. With AJAX or dynamic JS scripts updates should be fast. The intersection of updates of contacts could be done in a single API call with multiple parameters (server-side intersection of data.) Or more scalable but a bit slower by doing multiple calls to separate server clusters (could even be calls to fetch plain files on CDN containing latest updates, aka "the wall".) I've played a bit with zero-knowledge technologies (similar to Clipperz) and I think it's very doable. The big problem problem lies on images and video. Last year I played with HTML5/canvas with encryption but the CPU and delay is prohibitive since it requires to work on decompressed images pixel-by-pixel. It would be nice if the major browsers started giving encryption and compression primitives to JS. And even nicer if there was an API to manage images, audio and video. Chrome NaCl looks also very promising for this kind of disruptive technology. Of course, the non technical problems of getting traction with users and making it commercially viable are still there just like in any other kind of social network service.
philcrissman 16y ago
My reason isn't entirely the privacy concern, though, I just think their UX and some of their design/application decisions are _horrible_. It's an excruciating site to use if you want to do anything other than the two or three things they _want_ you to do. * Sharing information: Easy! * Removing information: Okay, not _too_ hard, but not as easy as sharing it! * Adding applications/pages: Easy! * Removing applications: Good luck doing it in less that four clicks, even assuming you know where to look! * Unliking pages: Where the heck... oh, there it is in little tiny letters way down the left column... And so on.
- steveklabnik 16y ago
  It's because their incentives aren't aligned with yours. Of course they want to make it easy to share, and hard to wall yourself off. I'm not sure how to properly align those incentives, but you can't really blame them.
  - philcrissman 16y ago
    That's true. But I'm not necessarily talking about walling myself off; it should be _just as simple_ to remove an application as to add it. Anyways, you're right, but still; if they're going to continue pursuing their own incentives and not their users, I predict they'll (eventually) lose their users.
pkulak 16y ago
Buzz. None of my friends are using it, however.
- philcrissman 16y ago
  I really wanted to like Buzz, but so far I haven't been able to really dig it. It's missing a certain je ne sais quois, but I don't know what it is.
ryanjkirk 16y ago
There is one actually, it's called spokt.com.

pg 16y ago

What features should it have? (Or not have?)

philcrissman 16y ago
I mentioned a few things that bug me in the above reply-to-myself... Should: - Be able to export all friend/contact information anytime. CSV is fine. - Be able to _easily_ block/remove applications & notifications. It's fine people can create "which star trek appliance are you?" quizzes, but let's make it _easy_ for these to be removed from my feed if I don't want to see them. - Maybe have a more twitter style follow where it does not need to be a mutual follow. Obviously less information is shared if I don't follow back; etc. Just pondering this, it might or might not be better. I actually don't think there's _that_ many things. There are a few key things that, if done right, would make it a better facebook. I'm just stabbing in the dark though. Can this post be considered a YC application? ;-)
- TrevorBramble 16y ago
  Better than removing the deluge of unwanted app notifications individually would be a list, off to the side, of apps your friends have used that you haven't opted in or out of yet. Opt in and they all show up, opt out and it's added to a list of unwanted apps in your profile that you can edit later if desired. Obviously this would make users happy but not app makers, because the way it currently works is free (obnoxious) advertising for them. The mixed-access to updates based on bi- or unidirectional association is one of the very nice things about LiveJournal's social scheme. I knew a few people who use FaceBook for one level of communication and retain their LiveJournal accounts for another, because of this.
ihodes 16y ago
Those basic ones allowing communication: essentially directed and non-directed asynchronous sharing. And those you 1) opt-in for 2) pay for. There's a long list I have, but that's a startup for another day. (Or for a time a month from now... when school lets out.)
davi 16y ago
1: I want ACLs meet Web. I want to publish personal stuff to close friends & family, professional stuff to colleagues, and random miscellany to the world, all via the same tool. I want to be able to customize the layout of my content depending on the audience. 2: I want this without my friends and family having to learn about OpenID or having to sign up for a new account somewhere. (I don't know how to do this, I am just saying what I want. Posterous got this part right.) 3: I want to know that the investment I'm making in generating content will last. This means I need to trust that the host of my content is not going to disappear, or get bought and change its behavior radically. Alternatively, it means that I am given the ability to migrate my content to a well-behaved, API-implementing host elsewhere (including, possibly, on my own server). Nothing I know of addresses this set of wants elegantly. Maybe because monetizing the social graph requires as much transparency from the participants as possible, at least according to current thinking. But I feel that eventually, what 'sharing on the web' means will inevitably evolve to accommodate this set of wants. If Facebook decided to be less evil they could do ACLs-meet-web pretty easily, since they already have such a huge installed user base. In fact, they may have already done this a long time ago for all I know; back when I canceled my account, they did not have this capability.
- gcr 16y ago
  You are describing the Internet. Self-run web sites, blogs, and instant messaging systems: 1. allow you to fine-tune who sees what, 2. do not require extra accounts (save comment systems, etc.), 3. are not tied to any one provider or company and thus will last for a long time.
  - davi 16y ago
    How do you see the internet providing built-in ACLs? http://en.wikipedia.org/wiki/Access_control_list http://en.wikipedia.org/wiki/Access_control_list Even if it did, your assertion is like equating assembly to a high level language. Just because you theoretically can do anything in assembly that you can in a high level language doesn't mean you would want to. I am trying to specify a set of attributes for a non-existent tool that would let me more fluidly publish what I write and create. ---- edit: if anybody reading this would care to explain why I am getting downvoted & the parent is getting upvoted so much, I would appreciate it.
codexon 16y ago
Another interesting question would be: how will it make money? Websites like Facebook and Myspace are shunning the simple advertisement model. Selling private data will become the new standard. A website similar to scope and feature like Facebook will probably not be able to sustain a positive cash flow or break even without selling people's data. UNLESS...they cut features like photo storage, or make a technological breakthrough. They are making ~$2/user/year (not including expenses). Since it recently supposedly became cash-flow positive, it would be reasonable to assume that the expenses are also around ~$2/user. The question then becomes, how much of this $2 is slashed off when you decide to not sell private data? You only have to look at the cost of conventional CPM ads (usually less than $0.01 per impression) and multiply it by how many times a Facebook user opens Facebook each year. My bet is that it is far less than 365/2 since Facebook claims 50% of "active" users login every day.
- nfnaaron 16y ago
  "Websites like Facebook and Myspace are shunning the simple advertisement model." ? Every one of my facebook pages (feed and profile) always has two or three prominent ads over on the right.
  - codexon 16y ago
    What I mean is they are selling your personal information __in addition__ to regular ads (which are selected by datamining your personal information). As opposed to just regular ads.
- kmavm 16y ago
  Can you provide an example of facebook "selling private data"? I work there, and it would be a great surprise to me if you could.
  - codexon 16y ago
    Sure. http://en.wikipedia.org/wiki/Facebook#Beacon http://en.wikipedia.org/wiki/Facebook#Beacon The prime example that resulted in a lawsuit settlement. I mean really? Opt-in to reveal your purchasing and browsing habits with selected companies and then spam your friends? http://gawker.com/5426176/facebooks-great-betrayal http://gawker.com/5426176/facebooks-great-betrayal Another opt-in that is equivalent to Facebook buying web coverage at the expense of user privacy. http://gigaom.com/2010/04/22/facebooks-instant-personalization-is-the-real-privacy-hairball/ http://gigaom.com/2010/04/22/facebooks-instant-personalizati... These websites are able to scrape your data to provide you a "personal" experience. Another opt-in. I'm not sure if FB is selling this api, but I'm willing to bet its a certainty. These might not literally be bundling up personal information and selling it outright, but given the track record here, I wouldn't be surprised if they did another "opt-in".
nfnaaron 16y ago
I think Google could do this entirely with blogger, maybe reader, maybe Picassa, and email, and a bit of code. It would be a huge boon to Android phones. Google groups is broken for some people, and I'd stay away from using it as a building block. Lots of people have negative opinions about Wave, me included. I don't even know if Buzz would help in any of this, the launch was so botched I never looked at it; people have told me it wouldn't work well for anything like this. Posterous or Tumblr could do it too. Google doesn't seem to be very motivated in integrating their various acquired apps; maybe this would be that motivation. Or maybe Posterous or Tumblr might find some motivation in this. It need not, maybe should not, and probably could not compete with Facebook. That's not a reason to not do it. Jones Soda does pretty well, yet I doubt if they think of themselves as a Coca Cola killer. Whether Google or someone else does it, I'd like: GENERAL - Total control. I don't want to be a control freak, I want to be a control super freak. I decide exactly what is private, shared and public. Everything that can be shared in any way should be obvious, and the settings interface for controlling privacy should be very obvious in: a) what setting controls what feature, b) what feature is controlled by what setting. If you have to guess how to change a feature's exposure, or what feature a setting controls, then you don't have control. - Long-lived settings. New features should never expose tomorrow what was not exposed yesterday. That's just so disrespectful on facebook's part. - New features should begin with the most restrictive privacy settings. - No apps or third parties. Make your money through advertising or paid features. If you're Google then you already have great advertising revenue and this just ties into it. - No friend can expose anything of mine to any greater degree than I have explicitly exposed it. DISCOVERABILITY - If I want to be invisible at any level, then I should be invisible as long as the rivers run and the winds blow, until I change it. Invisibility includes: - The outside world. I'm either discoverable to anyone outside the system, or I'm invisible, according to my settings. - People inside the system. I'm either discoverable inside the system, or I'm invisible. - My groups of friends. I'm either discoverable to my friends' groups, or I'm invisible. EXPOSURE - I control what parts of my data are exposed, and how far out. This is both independent of discoverability and cooperative with it. For example, no matter how far out I might expose a photo album, it can't be exposed farther out than my discoverability. On the other hand no object needs to be as exposed as my maximum discoverability. If I'm invisibile outside the system, then I can't set the exposure of a photo album to outside the system. If I need a photo album or something else to be exposed farther out than my discoverability, then I'd need to create that album outside the system (say, as a picassa album in a different, public account). I might be able to link to such an album from inside (and that link would have the same discoverability and exposure protections as any other object). FUNCTIONS - Discoverability. Anyone can discover me, as long as they're within one of my discoverability spheres. If I've set my discoverability to "world" then you can find me through search or browse. You can send a message to me if I've enabled that (with enough hurdles to discourage spam). If you're a member of the system you can request to friend me. If my discoverability is more restrictive, then you might only be able to discover me if you're already a member. Or you might be able to discover me only if you're a member of one of my friends' other groups. - Exposure. Objects that you post or share have a settable exposure level. Posting objects to a group gives those objects the exposure that you previously associated with that group. - Chat. - Alerts. If a friend logs in while I'm in, I want to know it, unless they've set that data to hidden. - The usual. Can post comments, pictures, music, books, links, "other stuff" and have conversations about them. Post things about yourself in your profile, with different parts of you profile's exposure individually controllable. Be alerted to comments in conversations related to you or your posts, and changes to posted objects. - A group of one. I should be my own group, similar to how some Linux distros create a group whose name is your login and whose membership is just you. I may want to keep track of things just for myself. I may want to stage new additions, initially to just me, and then exposed to a gradually wider audience as I see fit. I may want to pull some things in to private that were exposed. Everyone starts as a member of their own group of one ("self" ?), unexposed and undiscoverable. You have to explicitly make yourself more discoverable, and you have to explicitly expose any objects in your group of one. - Explicit groups of two or more. I should be able to explicitly band together with other people and form a group. Everyone in the group is aware of everyone else in the group. We can all post, converse and share. The group can be invisible or discoverable. I can expose my membership in the group to any group outsider who can see the group, or I can hide my membership from anyone outside the group. The email analogy to this might be cc:. - Implicit groups of two or more. I should be able, on my own, to place people and groups inside of implicit groups that only I am aware of. If I post to an implicit group, each individual or group member of an implicit group receives the post or other object, and can carry on a conversation or other activity related to the object, but no individual or group within one of my implicit groups is aware of any other members of the implicit group; they aren't even aware that they themselves are part of one of my implicit groups, although they might guess from my style of writing. The email analogy to this might be bcc:. The names of the two previous group types, explicit and implicit, may not be the best names. I initially described them as public and private, but those terms had too much confusing overlap with the notions of discoverability and exposure. - Some groups owned by individuals and some groups owned equally by members. Not sure about this. Worth thinking about. - Group union and intersection. I'm a little fuzzy on this one. Say you broadcast post an object to two groups. Person A is a member of one group, B is a member of the other group, yet the two people A and B have no relationship to each other within the system. If people A and B's discoverability settings would allow it, then they should see each other's comments related to the posted object and be able to take part in that object's conversations with each other as if they were both members of the same group. If their discoverability settings would not allow either to discover the other, then they should not see the others' comments. If this function were done poorly, then gaps in conversations related to an object might be disorienting. - Acknowledgement of the outside world. Under my control and choice, I should be able to make automatic tweets, or posts to my external blog, etc, based on what I post and where inside the system. I should also be able to make incoming posts from the outside, like ping.fm to twitter and facebook. - Data migration. I should be able to export part or all of my data quickly and easily. - Data ownership. I should be able to change the exposure of data that I post, and to delete it. - Existence ownership. If I leave, I should be able to totally delete myself. Or, if I want, I could de-activate myself. But it shouldn't take searching outside the system to find a little known method for deletion. If I delete posts or data that are part of conversations, the system should insert a placeholder, but if it's mine I get to kill it. IMPLEMENTATION For this part I'll assume Google is implementing it, because they already have all the parts and we're all familiar with them. Applies equally to Posterous, Tumblr or a player to be named later, acknowledging that they would have more to build. The fundamental activity in the system is posting objects and having conversations about them. An object might be text, media, a link, etc. Posting objects could be handled very well by blogger. You can put pictures and links in a post, and there's a comment mechanism. The comments would continue the conversation. You could also handle posts and reading with gmail, because it has excellent conversation threading. If you use gmail you'll have to be very careful about leakage, in or out. Slashdot's forum has excellent threading, for inspiration. Alerts and keeping up to date could be handled by logging in a lot and reading your stream, or via Google Reader, or by email/gmail, or any combination of them. Again, be very careful about leakage in and out. Be nice to be able to post via email/gmail (hi Posterous!). Media and albums by Picassa, but not just photos, anything a member has the right to post: books, audio, video. Implementing some of this might be a challenge. Pick a subset. Iterate. UI, UX, UXB If you find anything looking like any part of Facebook, you're probably doing it wrong. User interface should tend toward minimal. There should be zero or one way to do everything. UI elements should not suddenly disappear, to be found again after a frustrating search. Facebook recently made the logout link hard to find, burying it in the Account menu. Probably because they want to discourage you from logging out. Don't do stuff like that, it's disrespectful, and even if there isn't evil intent it encourages discussions of evil. You should be able to easily choose where anything will be posted, and not have to worry at all that you might not understand where it's going. When you view anything, you should get some indication of who else can see it, general by default but able to drill down to groups and individuals if you like. Name everything well, once, and use only that name, everywhere. You shouldn't need a help system. But someone will want it, and it should be worth reading. You should also be able to get to help for specific functions, from the context of that function, without having to abandon your context and go into "reading the manual" mode. There's probably more.
tmsh 16y ago
They shouldn't be so successful. They made the same mistake as Goldman. In this case, small, problematic privacy features (people being able to tag you without your permission, etc.) that weren't exhausted + ambitious new platforms and exponential success suddenly caused them to reach a critical mass where people think twice about their service. It's all about the integrity of the core feature set, and avoiding future, short-term successes that erode into it. Ironically, one only really notices the erosions (which, as with Apple and Goldman, are really not unique within their industry) after a certain amount of success.
alecco 16y ago
1 Reasonable privacy 2 Plausible deniability (like OTR) 3 Fast updates 4 Easily customizable 5 Client-side extensible interface (Greasemonkey style) About reasonable privacy, the service shouldn't be able to read content. The provider can tell Alice (user on IP 1) sent something to Bob (user on IP 2) but the message content itself shouldn't be compromised. In particular profile-identifying details should be reasonably safe. This requires client-side encryption, of course. Plausible deniability isn't that tricky, it mostly requires careful use of protocols at design phase and there is OTR to borrow. With AJAX or dynamic JS scripts updates should be fast. The intersection of updates of contacts could be done in a single API call with multiple parameters (server-side intersection of data.) Or more scalable but a bit slower by doing multiple calls to separate server clusters (could even be calls to fetch plain files on CDN containing latest updates, aka "the wall".) I've played a bit with zero-knowledge technologies (similar to Clipperz) and I think it's very doable. The big problem problem lies on images and video. Last year I played with HTML5/canvas with encryption but the CPU and delay is prohibitive since it requires to work on decompressed images pixel-by-pixel. It would be nice if the major browsers started giving encryption and compression primitives to JS. And even nicer if there was an API to manage images, audio and video. Chrome NaCl looks also very promising for this kind of disruptive technology. Of course, the non technical problems of getting traction with users and making it commercially viable are still there just like in any other kind of social network service.

philcrissman 16y ago

My reason isn't entirely the privacy concern, though, I just think their UX and some of their design/application decisions are _horrible_. It's an excruciating site to use if you want to do anything other than the two or three things they _want_ you to do. * Sharing information: Easy! * Removing information: Okay, not _too_ hard, but not as easy as sharing it! * Adding applications/pages: Easy! * Removing applications: Good luck doing it in less that four clicks, even assuming you know where to look! * Unliking pages: Where the heck... oh, there it is in little tiny letters way down the left column... And so on.

steveklabnik 16y ago
It's because their incentives aren't aligned with yours. Of course they want to make it easy to share, and hard to wall yourself off. I'm not sure how to properly align those incentives, but you can't really blame them.
- philcrissman 16y ago
  That's true. But I'm not necessarily talking about walling myself off; it should be _just as simple_ to remove an application as to add it. Anyways, you're right, but still; if they're going to continue pursuing their own incentives and not their users, I predict they'll (eventually) lose their users.

pkulak 16y ago

Buzz. None of my friends are using it, however.

philcrissman 16y ago
I really wanted to like Buzz, but so far I haven't been able to really dig it. It's missing a certain je ne sais quois, but I don't know what it is.

ryanjkirk 16y ago

There is one actually, it's called spokt.com.