Windley's Technometria - Latest Comments in Why Didn't PubSub Become Twitter?

Re: Why Didn't PubSub Become Twitter?

bobwyman — Wed, 18 Nov 2009 07:24:08 -0000

PubSub.com is live again -- under new management...
http://bob.wyman.us/main/20...

Re: Why Didn't PubSub Become Twitter?

BillSeitz — Thu, 14 May 2009 11:05:19 -0000

I suspect that SMS provided, along with novelty, a rationalization for the message-size-limit which drove the cultural acceptance of tiny posts. If you had tried to launch a microblogging service without SMS, I think you would have had too much resistance.

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Wed, 13 May 2009 22:03:22 -0000

Bob,

Sorry to have unfairly maligned you in any way. Maybe I wrote it
poorly, but I was talking about the vision of the community at large,
not PubSub or you specifically. Sometimes technology doesn't take
because even if the founders have the vision, the people they're
talking to can't understand it yet because they don't have the
conceptual building blocks.

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Wed, 13 May 2009 21:56:37 -0000

I probably should have. :-)

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Wed, 13 May 2009 21:56:35 -0000

Yes, I agree that SMS was an important factor originally and created
the novelty. Now SMS is just an afterthought.

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Wed, 13 May 2009 21:56:33 -0000

Twitter isn't two-way. It's one-way as well. That's what it means to
enable asymmetric following. There are plenty of people I follow who
don't follow me and people who follow me who I don't follow. That's
one-way. It only *looks* two-way when two people follow each other,
but it's just two one-way conduits.

Re: Why Didn't PubSub Become Twitter?

bobwyman — Wed, 13 May 2009 16:33:42 -0000

Phil, you wrote: [PubSub lacked] "Vision and concept—not technology." Certainly, we had the technology -- PubSub's prospective search matching engine would probably still be vastly better than anything that has been deployed since... Given the kind of traffic reported by microblogging services today, PubSub's engine would have been able to handle the traffic (including full-text "track") without much problem.

But, I think this is probably the first time in my career that I've ever been accused of not having "vision"... The problem was not a lack of vision or concept, rather, it was not seeing well enough how to get from here to there (and not having the right management team at PubSub). The most important part of what I couldn't see was how to convince others that what I was building was valuable. The problem was communicating the vision... You may remember that I spent a great deal of time trying to convince people that "real-time" search and real-time message routing was valuable -- but virtually everyone responded that the minutes to hours of latency you got from polling RSS files was sufficient for all known applications. They were wrong, but their being wrong didn't help us at PubSub since we didn't know how to convince them that we were right.

Not only did we spend a great deal of time at PubSub working on real-time and prospective search, we also worked on a great many other issues that would be necessary to deploy things like microblogging as a *protocol* rather than as a closed, proprietary service. Thus, our thinking about this kind of application went into the Atom Format (things like the atom:source field, etc.). Also, I spent a great deal of time first on BEEP/APEX and then later on XMPP and the XEP-0060 the XMPP PubSub specification so that we'd have "push" protocols optimized for real-time applications rather than just HTTP's pull (Push is needed to get "real-time"...). Our projects like FeedMesh and efforts related to pinging were intended to build the low-latency data exchange backbone that would be needed to tie together the eventually vast torrent of small posts coming from millions of individual publishers. StructuredBlogging was intended to begin the process of popularizing semantic tagging in order to enable new real-time applications based on the pubsub protocols...

A lot of PubSub's work is being leveraged by people today. Twitter used to publish Atom entries over XMPP and the Laconi.ca based systems do so today. Companies like Gnip are basically proprietary implementations of FeedMesh. And, the applications of microblogging are rapidly broadening to the point where we'll need to consider semantic tagging and/or rules for creating structure within messages... So, the stuff we worked on wasn't a complete waste -- it is being used today, as it was intended to be used. It's just a little later than I expected and we get no credit for any of it. Ah well...

But, back to the "vision" thing... I started working in this area back in the early 80's and my vision for it hasn't really changed very much since then. But, I decided then that there wasn't enough network infrastructure or pent-up need in the market to make it work then. So, I waited. Later, I was horrified during the mid-90's when everyone got excited about "Push" technologies (PointCast, etc.) since I thought I might have waited too long -- even though I was sure that PointCast and friends were too early. Finally, in 2000, I decided that the net and people's understanding of it was developed enough to support applications like PubSub and what you see today as microblogging. Clearly, even though I'd waited 16 years, I hadn't waited long enough...
The problem wasn't not having the vision, it was picking the wrong time, the wrong team and the wrong approach to make that vision a reality.

bob wyman

Re: Why Didn't PubSub Become Twitter?

dromescu — Wed, 13 May 2009 16:28:49 -0000

I agree with Chris, SMS - Short Message Service is the most use asynchron/synchron messaging in the world on mobile space. I used SMS more like 15 years for messaging, machine2machine communications, Alerts systems, advertising a.s.o
Why text messages are limited to 160 characters.Where is the Truth, here http://tr.im/kn1m or here: http://tinyurl.com/c5ko6c ...

Re: Why Didn't PubSub Become Twitter?

dave — Wed, 13 May 2009 14:56:20 -0000

Did you see the <cloud> element in RSS? It's been there since version 0.92.

Re: Why Didn't PubSub Become Twitter?

sull — Wed, 13 May 2009 13:11:11 -0000

my saying "unnecessary" was a poor choice of words. my thinking on this is that as we move forward, the data networks should improve while more people have phones equipped with publishing apps.
i also was broadly referring to the trend of twitter... that is now so big and not nec as centric to mobile devices as it was when it started as a status update service. now, we have a plethora of desktop apps and the website itself is improving.... and it's gone mainstream.
so, SMS is still an important component, especially for the reasons you mentioned. but a larger percentage of users will not use SMS on twitter.
when i said that "SMS is not a driving force anymore", i meant in the context of twitter specifically.

Re: Why Didn't PubSub Become Twitter?

Chris Messina — Wed, 13 May 2009 12:57:56 -0000

I completely, 100% disagree.

SMS is still the only universal API for getting data from any phone into web services... and furthermore, is the only API that stands up when WiFi/Edge/3G networks go down or experience service interruption.

It's also increasingly necessary from an international perspective (that is, I might tweet by SMS when roaming, but I'm certainly not going to roam on data).

While it's true that phones are becoming more capable as publishing tools, I think that the network is still a great limitation on what we're able to do with these devices.

Re: Why Didn't PubSub Become Twitter?

sull — Wed, 13 May 2009 12:31:53 -0000

at the same time, you could have just posted short messages to your blog in 2002 :) along with linkbacks of course. http://en.wikipedia.org/wik...

Re: Why Didn't PubSub Become Twitter?

sull — Wed, 13 May 2009 12:29:40 -0000

this is also why twitter et all should evolve to allow longer messages intuitively/seamlessly/logically.

Re: Why Didn't PubSub Become Twitter?

sull — Wed, 13 May 2009 12:28:22 -0000

and now the trend will be that SMS is unnecessary as phones have web apps installed to do direct client/server posts and email. SMS can continue to be useful for private texting to those who dont use these social services. but obviously, SMS is not a driving force anymore.

Re: Why Didn't PubSub Become Twitter?

LLiu — Wed, 13 May 2009 12:25:54 -0000

RSS is one-way. Twitter API is two-way. People prefer conversations more than just listening/reading.

RSS missed the boat by sticking with simplicity rather than provide a lightweight version of NNTP. To me, Twitter API is really not much more than a Web 2.0 version of NNTP.

Re: Why Didn't PubSub Become Twitter?

dave — Wed, 13 May 2009 11:53:19 -0000

I listened to Mitch Kapor's speech at GlueCon today and it left me bewildered. What is it about RSS that inspires such disrespect? Mitch speaks with great fondness of FTP and Gopher, but "RSS is Dead." Of course nothing Mitch says will have any effect on whether RSS is dead or not, it's a format for crying out loud, it never was alive. Mitch's hypothetical death for RSS says more about Mitch, RSS can't hear him. And btw, every twitstream is also available in RSS, so as Twitter grows, so does RSS. Mitch.

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Wed, 13 May 2009 11:51:50 -0000

I have to disagree. I think that the tools were all there and could have been used for exactly what Twitter has done. I don't see it as apples and oranges at all, but rather a matter of perception. You have decided what blogs are "wordier and more controlled discussions or conversations" but that doesn't mean they have to be. In fact, there doesn't even have to be a blog--just an RSS feed that people can "follow" and see an updates you make.

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Wed, 13 May 2009 11:49:07 -0000

But there's nothing about RSS that requires that. My point is that it was merely perception that kept us from using RSS for this. We thought of RSS in a particular way and that limited our thinking.

Re: Why Didn't PubSub Become Twitter?

rkt — Wed, 13 May 2009 10:22:15 -0000

Unfortunately rss feeds don't give you the kind of interactive experience one looks for in twitter. If we can get microblogging to allow comments without complicated auth requirements and without worrying too much about spam, then that could become a good alternative

Re: Why Didn't PubSub Become Twitter?

Chris Messina — Wed, 13 May 2009 09:10:02 -0000

Yeah, I'd agree that SMS penetration is what made Twitter. Laptops are too clunky to get out to post 140 characters — but using a phone is perfect — especially since everyone has one. Once we finally got to the point where everyone could use their phones to send text messages did the preconditions for Twitter to exist.

Oh, and it was no longer just a nerdy thing to use online social networks, likely thanks to MySpace making it culturally "cool" to have a web profile.

Re: Why Didn't PubSub Become Twitter?

Jesse Stay — Wed, 13 May 2009 00:16:44 -0000

I think SMS and Track via SMS is what did it for Twitter. Then, once they had the masses, they were capable of removing it down to exactly what we had before, but in more limited form.

Re: Why Didn't PubSub Become Twitter?

Phil Windley — Tue, 12 May 2009 23:55:23 -0000

Yes it did. And you've used blogging like Twitter forever--mixing
short posts with longer ones.

Re: Why Didn't PubSub Become Twitter?

dave — Tue, 12 May 2009 19:48:56 -0000

Radio 8 had all the tools, in one app, no cobbling needed, even a 1-click retweet, in 2002.

Re: Why Didn't PubSub Become Twitter?

Jesse Harris — Tue, 12 May 2009 17:34:19 -0000

While some people may use Twitter as a replacement for RSS or see that RSS can do what Twitter does, it's comparing apples and oranges. Twitter would be best described as a chat room where you get to filter what participants you want to hear from and respond anytime from immediately to days later. Blogs and RSS fulfill entirely different needs for wordier and more controlled discussions or conversations.

While one can do the job of the other, that doesn't make it a good tool. Just like using a flathead screwdriver instead of a crowbar when tearing down a wall.