Monday, December 27, 2010

Bandwidth is the new coverage

Something caught my ear on the radio today: an ad from a major cell provider twitting the competition for making all sorts of claims about speed (and then claiming to be the fastest themselves ... so there!)

Not so long ago the battle was over coverage, but evidently coverage is now about as good as it's going to get.  The internet connectivity of a phone is now a bigger selling point than the mere fact that you can call people with it.  If you're outside the US, or just more into smart phones than I am, this may well be old news.  I tend to use my phone predominately as a phone and an alarm clock, probably because I almost always have access to WiFi and a decent-sized screen and keyboard [Hmm ... this one might be worth a followup -- D.H. Dec 2015].

So far I've found a smartphone most useful as an easily portable GPS, an application for which I'd gladly trade speed for coverage.  A GPS function without a usable map is not much use if you're lost out in the boonies.  But maybe that's just me.

What also caught my ear about the ad was that carriers appear to be claiming to be "X times faster."  Faster than what, I'm not really sure, but it sounded reminiscent of internet providers claiming to be "X times faster" than dialup.  Even with that nice low bar set, providers had trouble providing quite what they promised.  Perhaps cell providers will do better in that regard?

Friday, December 17, 2010

Another fine timesink

This one's from Google Labs, as it turns out, though the radio piece I heard had it coming from Harvard.  According to the site itself, it's the product of "a team spanning the Cultural Observatory, Harvard, Encyclopaedia Britannica, the American Heritage Dictionary, and Google*."

So what is it?  It's culturomics.  What's that, you may ask?

There's a bit of hype, mostly in the press, about opening up a "new field", but the web site is simply a tool to mine the Google Books corpus for trends in word frequency.

OK, maybe that doesn't sound overly exciting, but try it. I'll be here when you get back.  [Looks like there's not much at that link anymore.  I think this eventually was supplanted by or morphed into ngrams, which is still alive and well -- D.H. Dec 2015]

Time permitting, I'll probably have a bit more to say about the word-mining itself on the other blog.

* As usual, I don't know any more about the Google end of it than you do [and I still don't -- D.H. Dec 2015] and if I did I would recuse myself. It does fit pretty squarely with Google's mission "to organize the world’s information and make it universally accessible and useful," though it's not clear how useful the site might be to the non-specialist.

Tuesday, December 14, 2010

It's a slow news day -- why not pick on Babelfish some more?

It's not necessarily a fair test of a translation engine's usefulness to take a short phrase and round-trip it from one language to another and back, but it sure is fun.  The good folks at Sporcle must have come to the same conclusion, judging by the "Babelfish Videogame" quiz.  Try it yourself.  See if you can figure out what "The Developmental Soccer which is Occupational" and "Spatial Aggression Person" started life as.

I got 7 out of 20, and that was still good enough 57th percentile, so garbled were the results.

And now for extra amusement, here's the above translated into, oh, say, Portuguese and back:

It' s not necessarily one has just tested of a translation engine' utility of s to take a short phrase and round-trip of a back language to another one and, but it is sure amusement. The good peoples in Sporcle must have come to the same conclusion, julg for " Babelfish Videogame" questionnaire. It tries it you yourselves. He sees if you can appear for you are that " The developing soccer that is Occupational" e " Space aggression Person" started life as.
I comeƧ 7 of 20, and that age 57th still good percentile sufficiently, truncated thus was the results.
E now for the extra amusement, here' of translated s above, oh in for example Portuguese and has broken back:

Sunday, December 12, 2010

@papabear this is @babybear. What's your 20?

I was listening to a piece on All Tech Considered about the hackathon Random Hacks of Kindness and was  duly impressed, not only by the presenter's brave effort to rescue the original meaning of "hack" from the dustbin, but of course by the whole idea of hacking together apps to make it easier to save lives and otherwise make the world a better place.  Well done, all.

One of the hacks was an app that would use Twitter traffic during disasters to help pinpoint where aid workers were needed most or could generally do the most good.  Again, very cool stuff.  Then it hit me: Twitter is the new citizen's band.  Think about it.  A populist medium allowing people to converse with strangers and broadcast to (a portion of) the world at large.  Users of the medium go by handles.  Traffic is subdivided into channels.  And, what led me to the conclusion in the first place, the traffic itself is a fascinating combination of pure drivel and vital information (with a fair bit in between).

Not that CB itself has gone away.  Not only is the technology still in use, evidently so is most of the slang I recall from the 70s or so.

Friday, December 3, 2010

How many ways can a person send a message these days?

Well, how many ways can a person send a message these days?
  • Talk to a person.
  • Join a group conversation
  • Write a letter to a person
  • Write a letter to the editor
  • Call a person on the phone
  • Call into a broadcast show
  • Participate in a conference call
  • SMS (text) on cell phone
  • Email
  • IM a person
  • Join an IRC or IM chat room
  • Update your status on your IM service
  • Send a message on a social networking site
  • Update your status on your social networking site
  • Tweet
  • Write a blog post
  • Comment on a blog post
  • Post a comment on a forum
  • Send a message in your virtual world or MMORPG
  • Produce a podcast
  • Participate in a webcast
  • Um, yikes, that's a lot, but it's not hard to come up with more.  Put a message in a bottle.  Update a bug report.  Post a sign.  Spray-paint your house.  Wear a message T-Shirt.  Put a bumper sticker on your car ...
I'm limiting (yes, limiting) myself here to things a typical individual might be expected to do, so writing a magazine article or holding a nationally televised press conference wouldn't count.  Even so, there's a seemingly limitless supply of ways to send a message.

When faced with such abundance, a plausible explanation is combinatorial explosion -- a reasonably small number of factors which can be combined in a large number of ways.  What are some possible factors?
  • Cardinality:  How many senders and how many receivers are there? The choices for each are one or more than one (denoted N) This is why I distinguished calling a person from joining a conference call or phoning into a show.
  • Symmetry: Are there specific roles, e.g., sender and receiver, or is everyone on an equal footing? 
  • Potential recipients:  Who could possibly receive the message, or perhaps better, what group do you need to belong to in order to be able to receive a message?
  • Potential senders: For the purposes of this exercise, this is generally "Anyone" (though in some cases it's narrower).
  • Access control: Who controls who can send or receive?
  • Persistence: Can the message be expected to be permanently available to the recipient?  This would be within the messaging system.  Once a message is received, the recipient can generally keep a private copy and/or resend it.
  • Latency: How much time typically passes between sending and receiving?  Latency can be higher or lower.  It can also be arbitrary, as in the case of forums and email.
  • Bandwidth: The maximum rate at which information can be transferred
  • Message size
  • Anonymity: Does the sender typically know who is receiving?  Does the receiver typically know who's sending? If so, can one easily make oneself anonymous?
So does that help sort anything out?  If it does, we should see a fairly wide variety of combinations, keeping in mind that some particular combinations may not make sense.  For example, 1:N cardinality implies asymmetry and persistence generally implies arbitrary latency.  Let's see what we've got (the table below is too wide for this layout, so I made it scrollable. Here's how)

MechanismCardinalitySymmetric?Potential receiversAccess controller(s)Persistent?BandwidthMessage sizeLatencyAnonymity
Talk to a person1:1YesAnyoneParticipantsNoHighArbitraryNegligibleNone
Group conversationN:NYesAnyoneParticipantsNoHighArbitraryNegligibleNone
Letter to person1:1NoAnyoneParticipantsYesLowSmallDaysPossible for sender
Letter to editor1:NNoSubscribersEditorYesLowSmallDaysPossible for sender
Call person1:1Yes, except at startAnyoneParticipantsNoMediumArbitraryGenerally negligible, enough to be annoying in some casesPossible for caller or receiver
Call show1:NNoListeners/viewersProducersTypicallyMediumGenerally limitedGenerally negligiblePossible for caller; audience is anonymous 
Conference callN:NDepends.  In some cases only some participants can talk.Depends.  Passcode may be requiredModeratorDependsMediumArbitraryGenerally negligiblePossible
SMS1:1NoService subscribersService providerYesLowSmallArbitraryTo the extent phone numbers can be anonymous
Email1:1 or 1:NNoAnyone with email (doesn't matter who provides the email service)No one, except that various providers may try to screen out spammersYesLow for typed text, can be high for large attached filesVaries, but generally at least megabytesArbitraryPossible
IM1:1Yes, except at startService participantsService providerOptionallyLow (again excepting file attachments)Low (again excepting file attachments)It's "instant" messaging, right?Possible
Chat roomN:NYes, except at startService participantsService provider (for access to service), moderator, in some cases (for access to room)OptionallyAs with IMAs with IMAs with IMPossible
IM Status update1:NNoService participantsService providerMay or may not be archived; persists until changedLowSmallArbitraryNone
Social network message1:1 or 1:NNoService participantsParticipants, service providerYesLowGenerally smallArbitraryNone
Social network status update1:NNoService participantsService providerAs with IM statusLowSmallArbitraryDefault state for readers
Tweet1:NNoAnyone with a web connection or cell phoneService provider (Twitter)YesLowSmallArbitraryPossible for sender, default state for readers
Blog post1:NNoAnyone with a web connectionBlog authorYesLowSmallishArbitraryPossible for author, default state for readers
Blog comment1:NNoAnyone with a web connectionBlog authorYesLowSmallishArbitraryPossible for author of comment
Forum comment1:NNoAnyone with a web connectionModeratorYesLowSmallishArbitraryPossible for author
Virtual world/MMORPG message1:NNoService participantsService providerNoLowSmallGenerally negligiblePlayers go by pseudonyms
Podcast1:NNoAnyone with a web connectionCreatorYesMediumLargishArbitraryPossible for creator, default state for audience
Webcast1:N or N:NDepends, as with conference callAnyone with a web connectionModeratorPossiblyLowLowGenerally negligiblePossible

One could argue over particular entries, some of the terms could be better defined, and I could add other factors, for example privacy, but it's pretty clear that combinatorial explosion is exactly what's going on.  There are a zillion different ways of sending messages because there are a zillion possible combinations of features one might like.