The Future is not Real-Time
notes.unwieldy.net
notes.unwieldy.net
It's probably mostly just my age (37), but I feel like these things come and go on a cycle.
My first brush with realtime was having a pager in high school. Wow neat, my friends can get ahold of me even when I am out in the streets! Huh, lame, all these people have my pager number now, and it beeps all the time and I never know if it is worth calling back.
Then came ICQ. Most of my friends and college classmates were all over that shit. I never understood it: text chat only seemed interesting to me late at night when I was too stoned (or whatever) to focus on anything. It combined the annoying immediacy of a phone call with 10% of the communications bandwidth of actual conversation.
Then I got a cell phone, but by that time we had caller ID and other people had cell phones (as opposed to the pay phones that had been calling my pager), and so it was only a minimal intrusion because just glancing at it told me that I didn't need to take that call right now (95% of the time).
So it wasn't too annoying, until the phones started to bleat our plaintive little demands to read the "text messages" that they had become capable of. I disabled the service.
Eventually we got to where we have data service and pocket computers everywhere. (I love that btw.) So now tons of people are pecking out little bleeps on facebook and twitter pretty much constantly. I don't really get this behavior any more than I got ICQ.
I like reading realtime twitter bleeps, say, when I search for something, or perhaps when I go on twitter to read stuff (if I did that), but why would I want them at other times?
I feel like Clifford Stoll, sometimes, ranting in my cabin in the woods about how e-commerce is never gonna catch on. Especially when somebody "pokes" me on facebook. (Thankfully, I only see that when I log on every month or so to see my sister's latest baby pictures.) But... what does that even mean? Why would I want people to be able to "poke" me? Whenever they want? And seemingly purposefully devoid of any content other thank the fact of poking me? Crazy!
So I fully admit that I don't get this latest round of 'realtime', but the solutions are still the same as the last times realtime came around. Just arrange to deal with information on your own time.
You want your machines to get data in realtime; sure, why not? And it will be cool if, in the future, we do indeed "find ways to artificially stem the constant flow of information through algorithmic summarization."
But there's a simpler, manual way to stem the constant flow of demands on your time: just reject them, and deal with them later, at your own convenience, or perhaps never.
What exactly is it that makes that hard for some people to do?
I personally get a lot of value from real-time information sources - links to new software releases or bug fixes, updates on social meetings with friends, uploads of photos from vacations and the like. I want this information, the trouble is (as the article says) I can't be available to filter and process it 24/7. So I end up reading through lots of junk to mine the value from the stream. If someone wrote an algorithm that knew what I valued and filtered / built digests for me, that would be great! Saying I should just 'deal with it later' doesn't really help though, because later there's more stuff to filter.
The only thing missing is a global "busy" status. I feel this is due to the lack of a globally unified text message (including IM and twitter) system. I don't think there's a polite way to kindly tell someone to urgently shut the hell up over text/IM, at least not that I'm aware of. The only annoyance I have with text communications systems is hearing "dingdingdingdingdingdingdingdingdingdingding" as someone sends me 15 messages in a row when I am unable to respond.
The future isn't real time. The past is. A phone call is. The future is sending information as quickly and painlessly as possible, and for "Gen Y", this includes text/IM. These services don't demand attention, beyond letting you know they're there. Then again, I grew up with AIM/MSN/ICQ, so again, I feel it might be a generational thing.
Quick one-offs by text are great. Voice is much easier in certain situations (when typing is difficult, when driving). And at times, voice allows for a much more rapid exchange than text. A quick 1-5 minute phone call can be far less interruptive than a 15 minute sequence of texts or IMs.
Sometimes, as well, it's simply nice, or the right thing, to talk to someone.
I feel no obligation to respond to and/or pick up any given communication from a stranger. Family, friends, and co-workers are generally given a much higher priority (though not absolute).
one is the difference between people being pushed around by realtime, vs. people having aggregators and agents consuming realtime - the human does not need to be glued to the screen to benefit from the feed. But if there is no realtime feed then this is not possible.
So if the slow web is about having control, then having access to realtime updates may be important to the slow web... odd as that might sound.
I view this the same way people view sports. People that watch sports want it real time because everyone discusses it as it happens (or immediately after). If the media you are consuming is important to discussions in approximate real time, then it will be important to receive that media as it happens.
This obviously doesn't fit everything, sports is just an example where it does.
"You should follow me on Twitter here."
Did you expect the people who invented Facebook and Twitter to turn off their cell phones?
Take, for example, Facebook Top News. The benefit isn't that it's not real time, it's just that the algorithm for deciding what's top news isn't driven primarily by the post timestamp. In the future defaulting to Top News makes sense, but there's no reason that feed shouldn't be real-time. If a close friend gets engaged, I would want a notification about that as soon as possible. But I want Facebook to be smart about when it interrupts me.
The same is true for Twitter's weekly digest. I'd love to have the tweets Twitter picks out in real-time rather than only at the end of the week. I may not want to be notified in real time, but I wouldn't mind if they went in a special list I could browse every day. Basically whatever metrics Twitter uses to decide what are the best tweets for me, I want those applied is real-time, and when a story meets the threshold, let me know.
I use Facebook's 'starred friend' feature and that's helped a lot to filtering which updates I want to see. It's still not nearly custom enough though - it would be nice if I could rank 'event types' and 'friend-to-friend connection strength' myself to help the filtering a lot more. I always feel as if the automatic filtering in Top News is too general (i.e. based on update views/number of replies), instead of being specific to me.
The concept of the Slow Web is catching on precisely because, as technology has matured, the idea of real-time communications has become the default, and as a result, using nearly any service exposes you to a bombardment of information. Technologically, it's easy: something happens, so you immediately fire off an email or trigger a push notification. Why wait? Why bother implementing a queue or a summary when it's so simple to just do it now?
Yes, users can always opt out, but often doing so means you either have to opt out of the service entirely (thereby losing the ability to benefit from it at all) or that the service falls out of your consciousness because they have no setting between "firehose" and "off", and there are so many other things demanding your attention.
Especially when services are ad-driven, they want to find ways to keep getting your attention. Every time they notify you of an update and you load the page, that's an ad impression. Every time you're prompted to share something and it gets a friend of yours to sign up, that's a new data point for advertisers. They have a vested interest in pulling your attention away from whatever you're doing and back to themselves, and real-time communication makes that easier.
You can choose to ignore notifications or put them off until later, but the fact that they're sitting there with unread information is often difficult to ignore. The fear of missing out drives us to keep checking; the fact that every fifth or eighth or twelfth tweet makes us laugh or think is partial reinforcement that drives us to try to read every one. Opting out through willpower alone is more difficult than it seems for most people.
Who says communication has to go over the web?
When I send a message or some other data to someone, like when I send something via FedEx, I'd like to know how long it's going to take to arrive. If it can be instantaneous, all the better. When the recipient chooses to read it is their business.
Real-time communications systems are quite valuable, in my opinion. Financial markets now rely on them.
But what I'd really like to see are more real-time operating systems for general use, by consumers. A real-time kernel makes computing very predictable. The time taken to complete a task can be accurately estimated before it is undertaken. Sometimes this is very useful. Sometimes the user might only want to do one or more single things at a time. Maybe there's a task I perform everyday and I want to be sure it will always take the same amount of time, each and every day.
In the real world, outside of computers, everyone knows "multitasking" is overrated. If you want results, it's better to be able to _focus_.
Maybe we might say the same for computers, in some situations.
I don't know, who?
When I send a message or some other data to someone, like when I send something via FedEx, I'd like to know how long it's going to take to arrive. If it can be instantaneous, all the better.
Why? Why do you care? As long as it arrives before the recipient can actually handle it, isn't it irrelevant?
Real-time communications systems are quite valuable, in my opinion. Financial markets now rely on them.
With disastrous consequences, some would claim.
Re:RTOS. The only thing it buys you is submillisecond scheduling, vs standard Linux millisecond scheduling. It doesn't mean the system is single-task, nor does it protect the task from being delayed by doing IO, which almost all tasks do (and if they connect to the internet, you can forget any kind of hard or soft realtime guarantees).
In the real world, outside of computers, everyone knows "multitasking" is overrated. If you want results, it's better to be able to _focus_.
For the most part, multitasking is not the problem. Application design and dependency on external systems is much more important. What's the point of having a single task running, if it's waiting e.g. for the DNS to resolve?
However, I believe the systems that do handle this communications data will need to be increasingly real-time. There will be so much data that unless it is indexed and collated in real-time, it will be impossible to perform queries to sort through mountains of un-indexed data. I think computers will become more like intelligent agents, bringing things that require immediate action to our attention and filing/organizing everything else for our later consumption.
There are already services, such as News.me and Prismatic, providing these services for Twitter. In the future, I expect the idea behind these services to be closer to the norm rather than the exception they are today.
> On a daily basis, we’re exposed to hundreds of articles, tweets, emails and advertisements <…> > We will find ways to artificially stem the constant flow of information through algorithmic summarization
It may be bad wording or misunderstanding, but I don't think we get exposed to it by some external force. There's a conscious decision one can make: which sources to check? how often? what to ignore?
I don't think the existence of technical tools and filtering algorithms would eliminate the need for these decisions.
I agree, and expect the retreat-from-realtime to be fed both by technological innovations -- better ways to reorganize, filter, and reprioritize -- and by a cultural maturation/backlash.
Those services that "simply want" our attention, and use interface patterns to trick us into giving them incrementally more of it, will be recognized as abusive. Even the traditional reverse-chronological timeline, by its mixing of wildly different (and often repetitive) messages, tends to falsely signal more importance/urgency/novelty than is actually present. It's been an easy and familiar bit of interface 'sugar', but in the wiser future we may consider it as tacky as tabloid front-pages and cable/headline-news teasers.