Social implications of filtering mechanisms

In response to my "Wearing my heart on my blog" post, marmelade27 wrote the following comment:

"You can use filters to make posts visible to only certain people or
you can make them completely private and not visible to anyone...
so make some custom filters and post away..."

I wrote out a nice, lengthy response to the comment, but LiveJournal apparently thinks I talk to much and wouldn't let me post it (it was over their allowed number of characters for a comment). So instead, here is my response as its own posting:


Ignoring, for the moment, how unintuitive the admin UI of LJ is, and therefore how much of a pain it would be to try and figure out how to use their filters, that still wouldn't really solve the fundamental problem.

You see, I don't really care if random strangers read my posts about who I'm pining over. And I don't really care if most of my friends read them, either. It's just when the people involved read them that it could get awkward.

And yes, I could put a filter dis-allowing just those people, assuming that I know their LJ username. But I'm assuming that the list of dis-allowed people is not shown on the blog. So if I write a post and disallow just the involved individuals, our mutual friends will see it and not know that the person involved is not seeing it. So the chances are fairly high that they'll ask the person "so what did you think about that post, eh?"

Ultimately, I guess it's analogous to talking about someone while they're in the room. Adding filters to dis-allow one person would be the equivalent of kicking that person out of the room before speaking, or otherwise distracting them. Everyone who does hear the comment doesn't necessarily know that the person is no longer in the room. And since the comment was made to the whole room, it's commonly assumed that what is being said is now common knowledge. This is why filtering by exclusion doesn't work very well. Or, at best, requires much more complex filters to work sufficiently.

This is why, in the blogging software I'm theoretically writing, filters are based on inclusion instead of (or in addition to) exclusion. Basically, there are three default permissions levels you can assign to a post:

  1. Public - Anyone can see it.
  2. Private - Only users who have logged in to your website can see it.
  3. Secret - No one sees it but you.

Public and Secret are all-or-nothing settings that cover 95% of the use cases. For Private, you've already done some inherent filtering by deciding who does and does not get an account on the site (equivalent of a friends list). In addition to that, you can create various user groups, and add those user groups to either a "show only to" or "do not show to" list (with do not show taking precedence, if the same person is on both lists). And when you make the post, both lists will be shown on the posting, similar to the "to" field in an email.

That way, if you create a usergroup called "confidential bitch session" and add just a few of your closest friends to it, they can see that the post is only being sent to a few people, and know not to spread the word to people who are not on that list. To return to our previous analogy, this would be the equivalent of grabbing a few people in the room and dragging them outside, to tell them something in private. They know that they're the only ones being told, so they know the information is private.

I believe that filtering based on inclusion instead of exclusion is much more similar to how people have already been conditioned to behaving with emails (as well as actual social interaction). If the email is to a widely distributed list, then the content is fair game. If there are only a few people in the "to" field, then you know it's not public info. So using a similar metaphor in your blogging filters would be the best way to insure that private info is kept private, and no one accidentally lets the cat out of the bag.

Furthermore, filtering by exclusion leads to far too much drama. If, for example, you were to specifically exclude someone from a party invitation, that person would likely be incredibly offended if they found out (and Murphy's Law tells us they *would* find out). It's a very passive-aggressive way of avoiding someone without being honest about it, and I don't like it when technology urges us to continue our societal bad habits.

In fact, the only reason that I would still allow filtering by exclusion is for the use cases where you want to exclude one person, and at the same time let everyone else know that you are excluding them. The obvious example would be when sending out an invitation to a surprise party. You could send it out to the "everybody" user group, but put "Joe Schmoe" in the exclude list, so that everyone except Joe gets the message. And right under the "to" field would be an "exclude" (or whatever - it would need a better name) field, which shows the list of excluded users to everyone who's reading the post. That way everyone knows that they're not supposed to mention the party info to Joe.

Wow. Was that a tangent, or what?

Anyway, since I'm pretty much stuck with LJ until I get off my arse and finish my own software, the conundrum remains. And even if I do finish my software, it'll be good for me to know the kind of problems that others are also having with the existing tools, so I can fix those problems to the best of my ability in my own tool. So getting to the root of the "what you say to whom" question is still going to persist, regardless of what technological toys we have to play with.

Don't you just love it when advances in technology make you go back and re-examine fundamental questions of social interaction? No? Maybe it's just me...