What
 · Technology
 · · Internet

The Fixing-JSON Conversation · Last week I sug­gest­ed some mod­est JSON im­prove­ments, and con­ver­sa­tion en­sued. Ob­vi­ous­ly, much was “He’s Wrong On The In­ter­net (again)” but lots was juicy and tasty, and worth con­sid­er­ing fur­ther ...
 
Fixing JSON · I’ve edit­ed a cou­ple of the JSON RFCs, and am work­ing on the de­sign of a fair­ly com­plex DSL, so I think I can claim to have dug deep­er in the JSON mines than most. We can eas­i­ly agree on what’s wrong with JSON, and I can’t help won­der­ing if it’d be worth fix­ing it ...
[33 comments]  
Inside Amazon · Gosh, it seems that my employer’s at-work cul­ture is the talk of the In­ter­net. Don’t know if I should share on the top­ic, but I feel the urge and blog­gers with the urge got­ta blog ...
[9 comments]  
Google Music · I use iTunes at home, and oth­er­wise Google Play mu­sic, which re­al­ly isn’t ter­ri­ble. Since the Net is echo­ing with screams about Apple’s cloud-music prob­lem­s, now might be a good time for a few words on the sub­jec­t ...
[3 comments]  
No Magic · I spent a cou­ple days this week in east­ern Wash­ing­ton State with a lot of se­nior Ama­zon en­gi­neer­s, all try­ing to dis­cern and in­scribe co­her­ent form on the roil­ing sur­face of global-scale cloud tech. This piece is here so I can pub­lish one pret­ty pic­ture and four wise words about what it means to be an en­gi­neer ...
[1 comment]  
451 Again · Back in 2012, fol­low­ing on work by oth­er­s, I sub­mit­ted an Internet-draft propos­ing a new HTTP sta­tus code, 451, to sig­nal le­gal block­age. As of to­day, the lat­est draft is a work item of the Hyper­text Trans­fer Pro­to­col Work­ing Group of the IETF. Doesn’t mean its fu­ture is clear, but I’m still hap­py. I’m post­ing here to fo­cus on one par­tic­u­lar turn of phrase, and to ask for a spe­cif­ic fla­vor of feed­back ...
[5 comments]  
RFC 7493: The I-JSON Message Format · The Olde ASCII is at rfc7493.txt. I’ll put a nicely-formatted HTML ver­sion here as soon as I pull a few pieces to­geth­er. This is re­al­ly, re­al­ly sim­ple stuff and should be about as controversy-free as an RFC can be ...
[13 comments]  
At Work · More re­portage from in­side the AWS fac­to­ry. Look­ing for leaks or mar­ket­ing? Nope ...
[2 comments]  
At Work · No, I’m not go­ing to be­come an “Amazon Blogger” (that ter­ri­to­ry is well-covered) but I have at-work sto­ries, nei­ther leaks nor mar­ket­ing ...
[3 comments]  
Keys in the Cloud · I just land­ed a nifty new fea­ture for OpenK­ey­chain. It’s sim­ple enough: If you want to com­mu­ni­cate pri­vate­ly with some­one, you need their key. So, just like when you’re look­ing for any­thing else, you type their name or email or what­ev­er in­to a search box and find it on the In­ter­net ...
[1 comment]  
Hangouts to Slack · I con­verse with peo­ple ev­ery day on all sorts of dif­fer­ent on­line chan­nel­s; I guess I’m sort of a chat con­nois­seur. Since I don’t have a work team these days, the on­ly big group chat that con­sumes any time at all is my lo­cal Ingress fac­tion com­mu­ni­ty. It’s been on Google+ Han­gouts for over a year, but they just moved over to Slack ...
[4 comments]  
Blow up the Cert Business · Check out Main­tain­ing dig­i­tal cer­tifi­cate se­cu­ri­ty by Adam Lan­g­ley over on the Google On­line Se­cu­ri­ty blog. Bad certs in the wild, many Win­dows users (but not on Fire­fox) vul­ner­a­ble. This is very, very bad. Let me elab­o­rate a bit and ex­plain how Google could solve this prob­lem ...
[8 comments]  
Privacy Levels · You should be able to ex­change mes­sages pri­vate­ly us­ing the In­ter­net. My pro­fes­sion should be work­ing on mak­ing this easy for ev­ery­one, in­clud­ing non-geek civil­ians who shouldn’t need to un­der­stand cryp­tog­ra­phy ...
[12 comments]  
Pervasive Monitoring Is an Attack · That’s the ti­tle of RFC 7258, al­so known as BCP 188 (where BCP stands for “Best Cur­rent Practice”); it rep­re­sents In­ter­net Engi­neer­ing Task Force con­sen­sus on the fact that many pow­er­ful well-funded en­ti­ties feel it is ap­pro­pri­ate to mon­i­tor people’s use of the Net, with­out telling those peo­ple. The con­sen­sus is: This mon­i­tor­ing is an at­tack and de­sign­ers of In­ter­net pro­to­cols must work to mit­i­gate it ...
[8 comments]  
Fat JSON · Most server-side APIs these days are JSON-over-HTTP. Devel­op­ers are gen­er­al­ly com­fy with this, but I no­tice when I look at the JSON that it’s of­ten, uh, what’s the tact­ful term these days? Let’s say “generously proportioned”. And I see clum­sy code be­ing writ­ten to walk through it. The op­tions for deal­ing with this are in­ter­est­ing ...
[28 comments]  
Ethical Privacy Choices · Here’s a lit­tle rant I post­ed to an IETF mail­ing list thread on whether the IETF should move its public-facing ser­vices to private-by-default mod­e. Some­one post­ed a re­ply sug­gest­ing that “the us­er gets to choose the de­gree of se­cu­ri­ty that they con­sid­er appropriate” ...
[2 comments]  
Counter-Surveillance · Surveil­lance on the In­ter­net is per­va­sive and well-funded; it con­sti­tutes a planetary-scale at­tack on peo­ple who need the Net. The IETF is grap­pling with the prob­lem but the right path for­ward isn’t clear ...
[11 comments]  
IETF 88 · I at­tend­ed to pitch in on JSON and OAuth work, and be­cause it was here in Van­cou­ver. But this meet­ing was re­al­ly about de­fend­ing the In­ter­net from those at­tack­ing it. Which is worth everyone’s at­ten­tion and de­serves more ex­pla­na­tion than I’ve seen in the main­stream me­di­a ...
[4 comments]  
HTTP Encryption Live-blog · The IETF HTTP Work­ing Group is in a spe­cial place right now. It held a meet­ing this morn­ing at IETF 88 on en­cryp­tion and pri­va­cy; the room was packed and, just pos­si­bly, nee­dles that mat­ter were moved ...
[10 comments]  
Security in Internet Protocols · IETF 88 is go­ing to be a pret­ty hot meet­ing, what with the world learn­ing about lots of ug­ly at­tacks against everyone’s pri­va­cy and se­cu­ri­ty. At the end of the day this is a pol­i­cy prob­lem not a tech­nol­o­gy prob­lem; but to the ex­tent that any­thing can be done at the tech­nol­o­gy lev­el, a lot of the peo­ple who can do it are here. So I think these dis­cus­sions mat­ter, and I’m go­ing to do some rare semi-live-blogging to re­lay in­ter­est­ing news as it de­vel­op­s ...
[4 comments]  
Editing JSON · As sort of a 2% pro­jec­t, I’m help­ing out over in the IETF, work­ing on a re­vi­sion of the JSON spec ...
[7 comments]  
Chromecast · It took like five min­utes to get it set up. It work­s. It’s great. It’s on­ly do­ing 5% of what it could ...
 
Good Internet Baseball · I was in OpenID meet­ings at Mi­crosoft all day Tues­day, and start­ed driv­ing home to Van­cou­ver at 4PM. This a fair­ly painful route at that time, but the Blue Jays and White Sox, via MLB on the Nexus 7, re­duced the pain con­sid­er­ably ...
[4 comments]  
JSON Lesson · I just learned (maybe ev­ery­one else al­ready knew) that it’s le­gal to have du­pli­cate keys in JSON tex­t. But please don’t ...
[5 comments]  
Private By Default · As of now, this blog’s pri­ma­ry ad­dress is https://www.t­bray.org/on­go­ing; note the red “s”. That means your com­mu­ni­ca­tion with it is pri­vate, which I think is the way the whole In­ter­net should be ...
[24 comments]  
How To Watch Blood and Chrome · We watched all of BSG re­dux, and even though we loathed Capri­ca thought Blood & Chrome de­served a shot. Plus it’s on­line. Easi­ly ac­ces­si­ble on my com­put­er, but I like to watch TV on my TV. Here’s how ...
[9 comments]  
On the Deadness of OAuth 2 · Wow, did Eran Ham­mer ev­er go off. His noisy slam­ming of the OAuth 2 door be­hind him has be­come a news sto­ry. I have opin­ions too ...
[8 comments]  
Need a Latin Legal Scholar · As some of you may have no­ticed, last week I pro­posed a new HTTP sta­tus code to sig­nal the sit­u­a­tion where a re­quest can’t be ser­viced for le­gal rea­son­s. Here­with the back sto­ry, and an ap­peal for le­gal help ...
[8 comments]  
Uncrippling Tablets · David Wein­berg­er points out in Will tablets al­ways make us non-social con­sumer­s? that tablets im­ply a less-interactive “lean-back” re­la­tion­ship with the In­ter­net and thus the world. It’s pos­si­ble we can fix this by just un­crip­pling the key­board; I can’t re­al­ly have a con­ver­sa­tion with the world if I can’t jam text in fast, in­ter­rupt­ing bursts of high-word-flow with flur­ries of lo­cal ed­it­s. I sure don’t want to live in a lean-back world.
[7 comments]  
On Privacy · Re­cent­ly, my em­ploy­er an­nounced that An­droid app developers’ pages come with a place to link to their pri­va­cy pol­i­cy. This is such an ob­vi­ous­ly, painful­ly good idea; I can’t pass up the op­por­tu­ni­ty for a short lec­ture on Why You Should Care and What You Should Do ...
[2 comments]  
4G Performance Silliness · I was scan­ning the mobile-tech news and saw a sto­ry on a per­for­mance shootout be­tween the LTE im­ple­men­ta­tions from Ver­i­zon and AT&T; I skipped by the link and can’t find it now, but that’s OK be­cause I’m here to de­bunk it ...
[12 comments]  
Telus Android Island Internet Win · For Net ac­cess from our cot­tage on Keats Is­land, we checked al­ter­na­tives and end­ed up get­ting a In­ter­net stick from Telus Mo­bil­i­ty, plug­ging it in­to my old Black­Book, and hav­ing that broad­cast WiFi. It worked, but not bril­liant­ly, with big la­ten­cy and reg­u­lar out­ages ...
[2 comments]  
50M · Which is to say, 50,000,000 bits per sec­ond to our home. It feels great ...
[15 comments]  
Anniversaries & Ideologies · I took my lit­tle girl to the Sun­day tod­dler drop-in at the lo­cal com­mu­ni­ty cen­ter and thought about this weekend’s birth­days: Wikipedia’s tenth and the IETF’s twenty-fifth ...
[5 comments]  
Obese Pipes · Jim Get­tys has been demon­strat­ing the se­ri­ous­ness of the “buffer bloat” prob­lem; see Home Router Puz­zle Piece Two — Fun with wire­less, and The crim­i­nal mas­ter­mind: bufferbloat! This is most­ly just to draw your at­ten­tion to Jim’s work, be­cause you can prob­a­bly im­prove your own In­ter­net ex­pe­ri­ence by act­ing on his ad­vice; but I have a re­lat­ed gripe of my own ...
[8 comments]  
Let Your Data Go · It’s like this: If you send da­ta to some­one over the Net, you can’t con­trol what they do with it. At least cost-effectively. Or, if you want a good out­come ...
[14 comments]  
Nothing Creative · Com­pared to my lap­top, the iPad lacks a key­board, soft­ware de­vel­op­ment tool­s, writers’ tool­s, photographers’ tool­s, a Web server, a cam­er­a, a use­ful row of con­nec­tors for dif­fer­ent sorts of wires, and the abil­i­ty to run what­ev­er soft­ware I choose. Com­pared to my An­droid phone, it lacks a phone, a cam­er­a, pock­etabil­i­ty, and the abil­i­ty to run what­ev­er soft­ware I choose. Com­pared to the iPad, my phone lacks book-reading ca­pa­bil­i­ty, per­for­mance, and screen real-estate. Com­pared to the iPad, my com­put­er lacks a touch in­ter­face and suf­fers from ex­ces­sive weight and bulk ...
[59 comments]  
On Namespaces · It so hap­pens that my name on is the front page of Names­paces in XML 1.0, a tech­nol­o­gy which is pret­ty broad­ly dis­liked. Wel­l, it seemed like a good idea at the time. But I think we’ve learned some use­ful things since then and can make some good con­sen­sus rec­om­men­da­tions for peo­ple do­ing this kind of thing, es­pe­cial­ly if they’re us­ing JSON ...
[5 comments]  
Tab Sweep — Tech · Here­with glean­ings from a cir­cle of brows­er tabs fac­ing in­ward at the world of tech­nol­o­gy. Some are weeks and weeks old: Am­ber Road, Clo­jure, tail re­cur­sion, cloud­zones, deep pack­et in­spec­tion, and key/­val­ue mi­crobench­mark­ing ...
[5 comments]  
Cloud Sweep · Since I’m work­ing on this stuff these days, a lot of cloud-flavored news drifts across my radar. Here’s some of it ...
[3 comments]  
Standard Cloud · Check out the DMTF Open Cloud Stan­dards In­cu­ba­tor (“DMTF” is the Distribut­ed Man­age­ment Task Force); Wil­liam Vam­benepe pro­vides use­ful back­ground in DMTF calls the ball on Cloud stan­dards. But wait, there’s more! OGF of­fi­cial­ly launch­es work­ing group to cre­ate an API for cloud in­fras­truc­ture ser­vices (“OGF” is the Open Grid Fo­rum). The right num­ber of com­pet­ing cloud-standardization ef­forts can’t pos­si­bly be two. Are you a lead­er in an ex­ist­ing stan­dards or­ga­ni­za­tion? Or would you like to start a new one? Now is the time to get in the game ...
[2 comments]  
Modeling Is Hard · In A post-mortem on the pre­vi­ous IT man­age­ment rev­o­lu­tion Wil­liam Vam­benepe writes, on the sub­ject of stan­dard­iza­tion: “The first les­son is that pro­to­cols are easy and mod­els are hard.” I agree about the rel­a­tive dif­fi­cul­ty, but think when it comes to in­ter­op­er­a­tion, pro­to­cols are very dif­fi­cult and shared mod­els usu­al­ly im­pos­si­ble. A cou­ple of ex­am­ples oc­cur to me ...
[9 comments]  
Cloudy IT · There’s a study out from McKin­sey, Clear­ing the Air on Cloud Com­put­ing; many read­ers will have seen the com­men­tary here & there around the We­b. I’d rec­om­mend tak­ing the time to page through the 34 well-put-together pages of the orig­i­nal. Its con­clu­sion is deeply wrong ...
[7 comments]  
Pumping Iron · I work at Sun be­cause I like com­put­er­s, so when­ev­er we an­nounce some, that’s a big day for me. Today’s iron is built around Ne­halem. There are a cou­ple of blades, a bunch of rack-mounts, plus 10GE and In­fini­band switch­es (I have seen more than one in­ter­nal Ethernet-vs-Infiniband fist­fight; juicy stuff for geek­s. My per­son­al bet is on Eth­er­net). There will be tons of press re­leas­es and so on start­ing at the Sun home­page linked above, but for the re­al poop you need blog­ger­s; the peo­ple who built these box­es and ran the bench­mark­s. Tushar Katar­ki has ag­gre­gat­ed them in Sun rise over Ne­halem. Let’s cut Tushar some slack on the cheer­lead­ing, he’s been work­ing on these for a while and has earned the right to a lit­tle rah-rah; and he links to re­al blog­ger­s ...
[6 comments]  
Sustaining the Internet · I spent Thurs­day evening and Fri­day at a meet­ing called “The Fu­ture of the In­ter­net and Sus­tain­able Devel­op­ment Consultation”, host­ed by the In­ter­na­tion­al In­sti­tute for Sus­tain­able Devel­op­ment (IISD), an ad­vi­so­ry or­ga­ni­za­tion fund­ed by var­i­ous bits and pieces of the Cana­di­an gov­ern­men­t. This par­tic­u­lar ex­er­cise is aimed at pro­duc­ing ad­vice for In­dus­try Cana­da, which has some rea­son­able con­cerns about the In­ter­net, Canada, and the world. Da­ta point: Canada, once sec­ond in the glob­al In­ter­net con­nec­tiv­i­ty score­card, is now nine­teenth and falling fast. Which isn’t good ...
[3 comments]  
C1 in NY · Com­mu­ni­tyOne, that is. Just a few weeks from now: March 18th and 19th in Man­hat­tan. I’ll be there talk­ing about my An­droid work, and I’m al­so work­ing very in­tense­ly on some oth­er things we may, with luck, show the world. The pro­gram looks gen­uine­ly in­ter­est­ing; I’ll be there all day to take it in.
[Up­date: Oh hey, as Si­mon points out, it’s free.]

 
Experimental Engineering · I spent some time look­ing through the just-emerging web OS tech­ni­cal doc­u­men­ta­tion, and it quick­ly be­came ap­par­ent that Palm’s ap­proach is rad­i­cal­ly dif­fer­ent from both Android’s and Apple’s. Since they’re all here at more or less the same time, run­ning the same Web brows­er on rough­ly equiv­a­lent hard­ware, this rep­re­sents an un­prece­dent­ed ex­per­i­ment in com­pet­i­tive software-engineering ap­proach­es ...
[17 comments]  
On Email · This con­ver­sa­tion launched when Gmail an­nounced its of­fline mode and has been swirling around: See Man­joo (sum­ma­ry: “Gmail is great!”), Mr­gan (sum­ma­ry: “No drag and drop, blecch”), and now Alex Payne’s The Prob­lem With Email Clients ...
[21 comments]  
What’s “Cloud Interop”? · I’ve seen ver­biage on this echo­ing around the Net while the var­i­ous Cloudy fes­tiv­i­ties go on down in the Bay Area. You could spend a lot of time par­ti­tion­ing and tax­on­o­miz­ing the in­terop prob­lem, but I’d rather think of it from the busi­ness point of view ...
[6 comments]  
Cloud Interop Session · I spent Tues­day at the Cloud In­terop event or­ga­nized by Steve O’Grady and David Ber­lind. Sci­en­tists say that even a neg­a­tive re­sult is use­ful in ad­vanc­ing knowl­edge; I’d go fur­ther and say that a wait-and-see at­ti­tude in the heat of a hype cy­cle is of­ten op­ti­mal. By those cri­te­ri­a, this was suc­cess­ful. My at­tendee count peaked at 51 ...
[3 comments]  
The Shape of the Cloud · There’s an in­ter­est­ing ar­gu­ment go­ing on about the business-structure fu­tures of the Big Cloud that ev­ery­one as­sumes is in our fu­ture. Some links in the chain: Hugh Ma­cleod, Tim O’Reilly, Nick Carr, and Tim again. A few points about this seem ob­vi­ous; among oth­er things, Ama­zon Web Ser­vices is re­mind­ing me pow­er­ful­ly of Al­tavis­ta ...
[9 comments]  
Lockin-Free Cloud? · Google’s DeWitt Clin­ton, in a com­ment on my Get In the Cloud piece, as­serts that both Google App Engine and Ama­zon EC2/S3 are al­ready lockin-free by my def­i­ni­tion. That’s not quite con­sis­tent with the word I’m hear­ing on the street. I’d ap­pre­ci­ate tes­ti­mo­ny and point­ers from oth­er­s, be­cause this is a re­al­ly im­por­tant is­sue ...
[14 comments]  
Mobility Blues · Th­ese days, I’m gloomi­er and gloomi­er about the prospects for the mo­bile In­ter­net; you know, the one you ac­cess through the sexy giz­mo in your pock­et, not the klunky old general-purpose com­put­er on your desk ...
[32 comments]  
New TLD Fun · I’m not sure whether this free-TLD idea is a good or bad thing in the big pic­ture, but you can have some idle fun think­ing ’em up; it’s al­most po­et­ic:
.geek, .n­erd, .asperg­ers
.van­cou­ver, .con­do, .rain
.dot, .sun, .web
.pr, .fan­boy, .whore
.plane, .train, .au­to
.good, .bad, .fail
.p­wn­s, .suck­s., .rocks
.rock, .jaz­z, .blues
.mo­jo, .mo­fo, .ho­mo
.school, .col­lege, .job
.n­ba, .e­pl, .afl
.bondage, .dis­ci­pline, .s­pank­ing
.fast, .s­low, .stop
.be­gin­ning, .mid­dle, .end

[11 comments]  
Not an OS · Last Thurs­day I tweet­ed: “I strive to main­tain an open mind when non­tech­ni­cal peo­ple talk about the ‘Internet OS’ or ‘Web OS’. Some­times it's tough.” I got some grum­bles by email and I think the sub­ject is worth more dis­cus­sion. Let me be more speci­fic: Nei­ther the In­ter­net nor the Web is much like an OS. And even if it were, that’d be the wrong way to think about what’s go­ing on right now ...
[24 comments]  
Online Data · That S3 out­age sure con­cen­trat­ed people’s mind­s. And al­most si­mul­ta­ne­ous­ly, EMC an­nounces that they’re get­ting in­to cloud stor­age. It’s ob­vi­ous to me that we’re nowhere near hav­ing worked out the eco­nomics and safe­ty and per­for­mance is­sues around where to put your data. There are some ar­eas of clar­i­ty; geek über-photog James Dun­can David­son, in The Eco­nomics of On­line Back­up, shows that for a per­son with a ton of per­son­al data, the on­line op­tion is re­al­ly unattrac­tive. And you do hear sot­to voce rum­bles about go­ing on­line in the geek hall­ways, for ex­am­ple “Amazon web ser­vices: 3x the price, 0.5x the re­li­a­bil­i­ty, and low­er scal­a­bil­i­ty than DYI. Buy on­ly for the low capex and lead time.” That’s from Stanislav Shalunov, who by the way is a damn fine Twit­ter­er. The big ques­tions re­main open.
[3 comments]  
Boingo · Prepar­ing for that Ju­ly trip to Ber­lin, I was wor­ried about get­ting on­line; Lau­ren poked around a bit and sug­gest­ed Boin­go. It was a huge suc­cess; I got an hour or two in at Heathrow go­ing and com­ing, and the lit­tle Ger­man ho­tel was al­so on Boin­go. I’d sort of thought, when we got back, that I’d can­cel, since it seems to be more a Euro­pean than US thing. But it keeps show­ing up here and there, prob­a­bly enough to be cost-effective. In par­tic­u­lar, O’Hare air­port, which has ong been T-Mobile (ex­pen­sive and slow) or noth­ing, now has a net­work that named “Concourse” which us­es Boin­go and is damn fast, I think maybe the fastest WiFi through­put up­load­ing pic­tures to on­go­ing I’ve ev­er seen. So if you’re trav­el­ing a lot, you might want to check it out.
[3 comments]  
Simple Anti-Spam · I read my non-Sun email through GMail these days. It’s amus­ing to watch the ebb and flow of spam, as the bad guys fig­ure out a way through the de­fens­es, then Google patch­es that hole; re­peat forever. But there could be a lot less. If I could pick the char­ac­ter sets I can read, then it’d be au­to­mat­ic that any­thing in Chi­nese or Cyril­lic or He­brew or Ara­bic or Far­si or De­vana­gari is spam. Plus, any­thing that men­tions lot­tery in the ti­tle or the body, even on­ce. Plus, any­thing where the ti­tle is of the form “From XXX XXX”. That’d catch at least 75% of what’s still get­ting through.
[14 comments]  
Names · Lau­ren has a prob­lem: it’s “about what to name the thing that names names”.
[3 comments]  
OpenID Patent Covenant · Sun just an­nounced a Pa­tent Non-assert Covenant on OpenID; chap­ter and verse and FAQ here. Si­mon Phipps has a use­ful write-up. But what re­al­ly im­press­es me is the text of the covenant it­self; four short para­graphs of sim­ple, al­most jargon-free, English. Why can’t we do this more of­ten? I’m told that our own Ed­uar­do Gu­tentag gets the cred­it. [Ed. note: I’ve been asked a cou­ple of times now why don’t do one of these for Atom, too. Good idea, I should have been work­ing on it and I’ve been pro­cras­ti­nat­ing.]
[3 comments]  
Tab Sweep · Most­ly technology-centric, this time ...
[4 comments]  
Neptune · To­day we an­nounced some­thing, and let’s stick with the code-name “Neptune” be­cause of­fi­cial­ly it’s called the Sun x8 Ex­press Dual 10 Gi­ga­bit Eth­er­net Fiber XFP Low Pro­file Adapter, which just rolls off the tongue, doesn’t it. I have to say it’s about the butchiest-looking NIC I’ve ev­er seen. Now, nor­mal­ly I would have by­passed the op­por­tu­ni­ty to blog, be­cause I’ve nev­er ac­tu­al­ly even seen a 10G wire or plug, and have no idea what the trade-offs are and what makes a good one. But our Ariel Hen­del wrote Rus­sian Dolls, the first para­graph of which dis­cuss­es love, soc­cer, and beer, and which in­cludes a re­mark­able pho­tograph; it is ac­tu­al­ly about Nep­tune. Speak­ing as a geek who likes fast net­works and good writ­ing, I think it’s won­der­ful.
[1 comment]  
Geek Fight · Ladies and gen­tle­men: in this cor­ner, in the Mav­er­icks col­ors, Mark Cuban! And in the oth­er cor­ner, in the geek thread­s, Bram Co­hen! Both fight­ers score, but this judge gives the first round to Bram. (And I think to my­self: what a won­der­ful world.)
[1 comment]  
On Email Clients · As I wrote re­cent­ly, Mac Mail lost some of its mem­o­ry in a crash and, when I re­stored things, stopped work­ing. I was keep­ing it run­ning so I could go search for things, and then one day I ac­ci­den­tal­ly clicked “Get new mail” and it did, so I guess it’d worked through its is­sues, what­ev­er they were. Now that all my non-IMAP mail is on GMail, I can switch clients back and forth a bit; so here­with, com­par­a­tive re­marks on GMail, Thun­der­bird, and Mail.ap­p ...
[25 comments]  
On Email · For years, I’ve had two email ad­dress­es; the cur­rent job and the long-term per­son­al one. The lat­ter is un­for­tu­nate­ly one of the world’s most pub­lic, ap­pear­ing among oth­er places on the front of the XML spec­i­fi­ca­tion, and thus gets a lot of spam. I mean re­al­ly a lot. Which was caus­ing some pret­ty se­vere pain, but I’m us­ing the Gmail dodge, and that helps quite a bit ...
[24 comments]  
Tab Sweep · Uni­fy­ing the­me: none. Item: Ex­cel­lent Rails-vs.-Django study. No axe to grind, ap­par­ent­ly. No ob­vi­ous win­ner, which is news giv­en the Rails hy­pe. Item: Dana Blankenhorn’s Means and ends in open source; very thought-provoking. My guess is that the im­mense li­cens­ing fees driv­ing the bloat­ed sales in­fras­truc­tures at Or­a­cle, SAP, and friends are small in re­la­tion to the whole soft­ware ac­quire/de­ploy/­main­tain mon­e­tary pie, so the size of the whole in­dus­try isn’t like­ly to change that much. Item: Irv­ing Wladawsky-Berger, grand IBM tech­nol­o­gy poo-bah, spec­u­lates about the fu­ture of the 3-D Web in An Unusu­al Meet­ing. Speak­ing as one who’s made two con­cert­ed ef­forts to build a 3-D rep­re­sen­ta­tion of the We­b, I sure hope he’s right. Item: I can read Takashi’s cat’s mind. He’s 100% fo­cused on how he can get in be­tween Takashi and the com­put­er. (Takashi’s amus­ing post is about “Engineer's 2.0 day-life in the midafternoon”.) Item: From Clay Shirky, So­cial Fact­s, Ex­per­tise, Ci­ti­zendi­um, and Carr; a care­ful, level-headed thought piece on what it means to be an ex­pert, in the con­text of Wikipedia and Ci­ti­zendi­um. Item: From “jbischke” at Learn Out Loud, a handy list of The Top 10 Ar­gu­ments Against DRM; we al­ready knew most of this stuff, but it’s use­ful to have it pulled to­geth­er, well-argued and in one place. Item: Everyone’s blog­ging Test your mu­si­cal skills in 6 min­utes!; I on­ly got 72.2%, sigh.
[11 comments]  
Telepresence · Check out Scoble on Cisco’s HD telep­res­ence. I think he’s un­der­es­ti­mat­ing the im­pact. I’m on the road ev­ery mon­th, some­times two or three times, and I hate the air­lines and the whole trav­el sys­tem with the heat of a thou­sand sun­s. Plus, I prob­a­bly cost Sun the best part of $100K a year in trav­el ex­pens­es. Sup­pose you have a pop­u­la­tion of em­ploy­ees like me; cut­ting trav­el costs in half for six of us would cov­er the $300K price point. Plus, think of the ex­tra ef­fi­cien­cy we’d get from not spend­ing all those end­less hours in lines at air­port­s. Of the trav­el I do, some part is ac­tu­al­ly need-to-be-there stuff; but if the HD telep­res­ence prices can be brought down a lit­tle and the vir­tu­al meet­ings cranked up a lit­tle, this stuff is not that far from hav­ing slam-dunk ROI.
[7 comments]  
Practical Transparency · A few days ago, our CEO Jonathan Schwartz sent a let­ter to SEC Chair­man Christo­pher Cox call­ing for SEC financial-disclosure reg­u­la­tions to al­low for pub­lish­ing ma­te­ri­al fi­nan­cials on the We­b. It’s ob­vi­ous­ly a good idea, but there are some im­ple­men­ta­tion is­sues. (Hey, I’m an en­gi­neer, I can’t help it.) ...
[12 comments]  
Back to the Land · Every year I spend a few days on a fam­i­ly member’s farm in Saskatchewan; this is usu­al­ly fol­lowed by many pho­tos of cows and Prairie land­scapes here on on­go­ing. The In­ter­net there is dial-up but us­able, with the ap­pli­ca­tion of some dis­ci­pline. On­ly this year, I fired up my recently-repaired Mac and it told me that my in­ter­nal mo­dem did not ex­ist. Re­peat­ed re­boots failed to make any pro­gress, and I couldn’t re­al­ly get to the Ap­ple knowl­edge base. That’s OK, all I re­al­ly need is email and I could use the secure-webmail in­ter­face for that. On­ly for some rea­son, that wasn’t work­ing; in­com­pre­hen­si­ble er­ror mes­sages about some­thing be­ing mis­con­fig­ured. Ouch; so I filed a trou­ble tick­et and called my boss’ ad­min and asked her to drop me a voice­mail if any­thing super-important went across our group alias. Then when I called back to check my mes­sages, my per­son­al­ized Sun 1-877 num­ber gave me a busy sig­nal. Clear­ly, some heav­en­ly pow­er did not want me in­ter­act­ing with work. So I closed the com­put­er with an em­phat­ic “snap” and went to take a crap, and the toi­let backed up.
 
SavaJe Shakedown · At Ja­va One, I pur­chased the “conference gadget”, a SavaJe Jasper S20. Since it sup­ports most of Ja­va SE, I thought I’d check out whether a com­pe­tent Ja­va de­vel­op­er who knows noth­ing about mo­bile is­sues could make it do any­thing in­ter­est­ing. It’s kin­da cute, and it’s an OK phone on my lo­cal GSM net­work. How­ev­er, the soft­ware de­vel­op­ment pack runs on­ly on Win­dows. It’s got a cam­era and shoots video, but when I plugged the USB in­to my Mac, the Mac said “unable to use that disk” and then crashed a few sec­onds lat­er. I have nev­er had a USB de­vice fail like this on any com­put­er be­fore. Al­so, the phone’s pref­er­ences menus are lame and lim­it­ed, so on the screen, I can’t keep the net­work op­er­a­tor name from over-writing the Date/Time dis­play. Fi­nal­ly, my net­work provider re­quires some slight­ly odd GPRS set­tings (e.g. fun­ny DNS port num­ber­s) and the phone doesn’t have a way to set those, so I can’t con­nec­t. I re­port­ed this on the SavaJe de­vel­op­er fo­rum­s, and some badly-programmed bot pre­tend­ing to be a “SavaJe Devel­op­er Ser­vices Engineer” mind­less­ly point­ed me at the di­rec­tions to the phone’s GPRS set­up screen, when I’d specif­i­cal­ly stat­ed that that screen didn’t do what I need­ed. SavaJe might turn out to be a good idea, but they’re do­ing their best to cov­er that up. [Up­date: As of some­time over the week­end, an ac­tu­al hu­man seems to be hav­ing a look at the sit­u­a­tion, over at the SavaJe de­vel­op­er fo­rums. So let’s see...] [Up­date: Got a nice email from Lar­ry Kaye, Man­ag­er of Devel­op­er Ser­vices over at SavaJe. Very to-the-point; yes, they know about most of my is­sues, are work­ing on them. That’s all any­one can ask for in a bleeding-edge developer-release pro­duc­t. Will re­port fur­ther.]
 
Multicast · I’m pol­ish­ing up Si­grid so I can write about it and (if anyone’s in­ter­est­ed) re­lease it; I switched the lat­est JXTA (2.3.7) in—it lives at the bot­tom lev­el so that the grid nodes and clients can dis­cov­er each other—and ev­ery­thing broke, don’t know why. I still think JXTA is the right an­swer down the road, but on im­pulse I cracked open Ja­va in a Nut­shell and Stevens’ Unix Net­work Pro­gram­ming, and poked around the on­line Ja­va tu­to­ri­als to read about data­grams and mul­ti­cast. Dam­n, mul­ti­cas­t­ing through the Ja­va API sure seems easy. That same day I whacked to­geth­er a cou­ple hun­dred lines of code to do sim­ple dis­cov­ery, with ag­ing even (so that you for­get about things you haven’t heard from in a while) and test­ed it on a bunch of com­put­ers with­in easy reach; and it all seemed to Just Work. There are lots of useful-looking knobs on the side to deal with TTL and NIC se­lec­tion and so, but the de­faults seem well-chosen. As long as you’re pre­pared to deal with the fact that any giv­en mes­sage might not get through, and to spend a cou­ple of min­utes here, it’s re­al­ly not very tax­ing. I’m start­ing to won­der why you’d need any lay­ers of ab­strac­tion at al­l. Ques­tion: Do Python/Ruby/Per­l/Javascript have these coo­lio easy-to-use li­braries too? [Up­date: Ja­son Brig­gs shows how to do it in Python (but it takes as much code as Java, that can’t be right.) Kevin Hamil­ton wrote to point out Spread, which seems to have multi-language sup­port. Arkaitz Bi­tori­ka rec­om­mends ze­ro­conf.]
 
How to Send Data · I’m sure this has long since dawned on all the re­al­ly smart peo­ple, but it’s re­al­ly start­ing to be­come fifty-foot-letters-of-fire-in-the-sky ob­vi­ous: if you’re send­ing any­thing across the Net, why would you ev­er send it un­com­pressed? And if you’re send­ing some­thing to any au­di­ence oth­er than the whole world, why would you ev­er send it un­en­crypt­ed?
 
On the EFF, Email, and Syndication · Paul Hoffman’s The fu­ture of some email may not use email is a short but closely-argued piece which is fair­ly harsh to both AOL and the EFF, and says smart things about email and syn­di­ca­tion. Worth a read.
 
Fat Pipe, etc. · I gath­er that David Isen gave a rous­ing talk at the re­cent O’Reilly Emerg­ing Tele­pho­ny Con­fer­ence, which in­volved chant­ing a lit­tle mantra about how net­work providers should be­have: Fat Pipe, Al­ways On, Get Out of the Way! David gen­er­ous­ly cred­its me with in­vent­ing the phrase, which is true, see Fast and Al­ways On, an otherwise-forgotten frag­ment from March 2003. David’s per­for­mance art got no­ticed. I’m not en­tire­ly a fan of David’s po­et­ry, but the Free­dom to Con­nect event that he was pro­mot­ing looks darn in­ter­est­ing.
 
WaPo Still Screwing Up · The Wash­ing­ton Post has no­ticed the brew­ing storm over the two-tier In­ter­net; Doc Searls gets the long-term cred­it for start­ing the storm brew­ing. The WaPo piece, The Com­ing Tug of War Over the In­ter­net, is clue­less, at a triv­ial lev­el in al­leg­ing that the de­bate is hap­pen­ing “on ob­scure blogs”, but most of all in the out­ra­geous claim that “Companies like Google and Ya­hoo pay some fees to con­nect to their servers to the In­ter­net, but AT&T will col­lect lit­tle if any ad­di­tion­al rev­enue when Ya­hoo starts of­fer­ing new fea­tures that take up lots of band­width on the In­ter­net. When Yahoo’s mil­lions of cus­tomers down­load huge blocks of video or play com­plex video games, AT&T ends up car­ry­ing that in­creased dig­i­tal traf­fic with­out ad­di­tion­al fi­nan­cial compensation.” While the de­tails of the deals by which the big boys buy band­width are closely-guarded se­cret­s, the no­tion that any of them can dra­mat­i­cal­ly in­crease their net traf­fic with­out pay­ing for it, that no­tion is just wack­o. I’ve sent Christo­pher Stern, the au­thor, an email, but this silli­ness is al­ready on the streets of Wash­ing­ton in a few hun­dred thou­sand dead-tree in­stances. The dif­fer­ence be­tween blogs and the main­stream me­dia is that when we screw up, we can most­ly re­pair the dam­age. [Up­date: Five days lat­er. No an­swer to my po­lite, friend­ly email. No change in the ar­ti­cle. So this “journalism” thing... it’s a pro­fes­sion where you can just make ran­dom shit up and print it whether it’s right or wrong, and ig­nore feed­back, and you just don’t do those “retraction” or “update” or “apology” things? Seems like a flawed, short-lived busi­ness mod­el to me.]
 
HBR1 · I cou­ple of years back I wrote about the ad­van­tages of us­ing iTunes or XMMS to soak up Am­bi­ent In­ter­net Brain Goo. Th­ese last cou­ple of days I’ve been learn­ing how to par­ti­tion the disks and in­stall mul­ti­ple op­er­at­ing sys­tem­s, some of ’em pret­ty bleeding-edge, on my Ul­tra 20, which in­volves quite a bit of wait­ing and ir­ri­ta­tion. I owe my con­tin­ued san­i­ty to HBR1; as far as I’m con­cerned they’ve got the best brain goo go­ing.
 
Internet Stamps · This week’s splogstorm and the end­less flood of email spam are two symp­toms of the same dis­ease. When you al­low peo­ple to add con­tent to the Net for free, the eco­nom­ic in­cen­tives to fill all the avail­able space with with spam are ir­re­sistible, and fight­ing back is dif­fi­cult, maybe im­pos­si­ble. This works be­cause, while the pay­off per unit of spam is low, the cost is ze­ro. Wel­l, we can solve all these prob­lems at on­ce. It wouldn’t be free, but it would be cheap and it wouldn’t be that hard. It’s called “Internet Stamps” ...
 
GAIM & Adium, Good Stuff · I suc­cumbed to peer-group pres­sure and gave Adi­um a try; it’s a Mac instant-messaging clien­t, wrapped around the Gaim code, that sup­ports more or less ev­ery IM sys­tem on the plan­et. I’m very im­pressed, it’s good stuff. Gaim is very solid, and as for the OS X wrap­ping... well, iChat makes na­tive AOL look klunky and prim­i­tive, and Adi­um makes iChat look slop­py and space-inefficient; it gets the job done and gets out of the way. Al­so, if you want to whis­per se­cret­s, it sup­ports OTR en­cryp­tion. I can’t push pic­tures through it or do video chat, so iChat will still get used some­times. But, as of now, you can talk to me on AIM, Jab­ber, Ya­hoo!, and prob­a­bly even MSN pret­ty soon. [Up­date: MSN too, tim dot bray at sun dot com (uh, do I know any­one on MSN?); I’m now su­pe­rul­tra­mega­con­nect­ed.] Adi­um: high­ly rec­om­mend­ed. [Up­date: There’s an ir­ri­tat­ing bug, Adi­um from time to time makes it look like I’ve dropped off AIM even though I’m still there. I’ll have to go poke around and see if the de­vel­op­ers know about this.]
 
An Echoing Silence · This week­end, I’m be­ing re­con­fig­ured; to be pre­cise, the vast crowd­ed com­mons of Sun email is all be­ing re­con­fig­ured, and this is my week­end to leap the gap from be­fore to af­ter. So my Sun email is in­op­er­a­tive till Mon­day or so; if you need to con­tact me elec­tron­i­cal­ly, a quick Y! or G search will turn up an­oth­er per­fect­ly good ad­dress.
 
Google Talk · Yep, it works with iChat. Yep, I’m signed on (my gmail ad­dress should not be hard to de­duce). Yep, XMPP as a pro­to­col prob­a­bly has the legs to beat any oth­er sin­gle can­di­date. Yep, it’s dumb that the chat ser­vices don’t in­ter­op­er­ate. Nope, I don’t nec­es­sar­i­ly trust G’s benevolance, but yep, it’s good not to have all my eggs in an AOL bas­ket. [Up­date: fur­ther thought­s.]. Item: Uh, any­one can run a Jab­ber server, is this news sto­ry no more than “Google runs one too?” Item: Ars Tech­ni­ca is dis­mis­sive. Item: Hey, Ap­ple; why do I have to have two bud­dy list­s, one for “classic” iChat and one for Jab­ber? Item: All these names I don’t know are ping­ing me, want­ing to add me as bud­dies. For the first few, I ac­cept­ed, then sent a mes­sage “Do I know you?” So far, nobody’s an­swered, so I re­moved them. [Up­date again: In­ter­op­er­a­tion be­tween iChat and any non-Mac GoogleJab­ber client is shaky-to-absent. I tried to use Nitro, a na­tive Jab­ber clien­t, and it went­up in smoke.]
 
Down & Out in Tulsa · If it seemed like on­go­ing wasn’t there, it wasn’t. My host has a dire tale of cas­cad­ing fail­ure, but the weird thing is, when the site was down, tracer­oute showed the pack­ets cas­cad­ing in­to Wil­liams Com­mu­ni­ca­tion, where they strug­gled from Pa­lo Al­to to San­ta Clara to Den­ver to Dal­las to dusty death in Tul­sa, Ok­la­homa. How un­for­tu­nate.
 
Holy Gigabytes, Batman! · In April, on­go­ing served a to­tal of 84.5 gi­ga­bytes of da­ta to the world. Par­don me, but that just seems to­tal­ly mind-bogglingly ridicu­lous.
 
First Skype · So I just made my first Skype cal­l, from a hall­way in Makuhari to Lauren’s cell­phone in Van­cou­ver. She was busy, so it was just “Hi, talk again later.” But it worked and it cost me €0.02, which seems pret­ty fair.
 
Stronger Foundations · The Net to­day is stand­ing on slight­ly firmer foun­da­tions than it was a few weeks back. On De­cem­ber 15th, the W3C is­sued Ar­chi­tec­ture of the World Wide We­b, Vol­ume One as a “Recommendation”, i.e. as close as they get to call­ing any­thing a stan­dard. If you read it, there’s a lot of fo­cus on URIs (Uni­form Re­source Iden­ti­fier­s, as in what we usu­al­ly con­fuse with the Univer­sal Repub­lic of Love). To­day saw the fi­nal pub­li­ca­tion of the URI spec­i­fi­ca­tion as RFC3986, al­so known (even more im­pres­sive­ly) as IETF In­ter­net Stan­dard #66; some of the IETF’s most widely-deployed pro­to­cols nev­er make it to “Full Standard” sta­tus. I’m hope­less­ly bi­ased be­cause I helped write both of these doc­u­ments, but I think that each is in­di­vid­u­al­ly im­por­tant and that the com­bi­na­tion is re­al­ly im­por­tan­t. Most peo­ple can’t pos­si­bly imag­ine how much work it is to grind these things out, and how many re­vi­sions it takes to get one right. I orig­i­nal­ly wrote all of what is now Sec­tion 6 of RFC3986, and I think there are very few sen­tences in there that haven’t been up­dat­ed a few times, al­most all for the bet­ter, and al­most all the work was done by Roy Field­ing, who de­serves a deep bow from us al­l. The In­ter­net and the Web aren’t things and they’re not places, they’re a mesh of agree­ments that al­low us to talk to each oth­er and to be less stupid. Th­ese two doc­u­ments are im­por­tant links in that mesh.
 
Post-and-Poll · The ar­rival and insanely-fast growth of syn­di­ca­tion/RSS tech­nol­o­gy brings a New Thing to the In­ter­net. Un­til re­cent­ly, there was on­ly one mes­sag­ing ar­chi­tec­ture known to work at In­ter­net scale. That was store-and-forward, as in how email work­s: I send my mes­sage to a com­put­er near me, which stores it and sends it to an­oth­er com­put­er near you (with re­tries and so on as ap­pro­pri­ate), which stores it, and you re­trieve it from that com­put­er. Syn­di­ca­tion has proven that a dif­fer­ent mod­el, post-and-poll, scales up too. It works like this: I post some da­ta which con­tains ei­ther mes­sages or mes­sage point­ers to a pub­lic place, and you poll pe­ri­od­i­cal­ly to see what’s new. The dif­fer­ence is that store-and-forward sup­ports anybody-to-anybody mes­sage traf­fic, while post-and-poll as­sumes sep­a­rate com­mu­ni­ties of senders and re­ceiver­s. Peo­ple who are de­sign­ing mes­sage in­ter­change frame­works that might need to be­come Internet-scale should con­sid­er this, and be care­ful of ar­chi­tec­tures that don’t fall in­to one of these two bas­ket­s, be­cause noth­ing else has yet been shown to work. [Up­date: Some­how I for­got to cred­it Mark Hap­n­er of Sun who point­ed all this out to me, pre­sent­ing it as some­thing ob­vi­ous, not a dis­cov­ery; but then I re­al­ized that it hadn’t been ob­vi­ous till he did so.]
 
Jamaville · I’m on va­ca­tion for a week in Saskatchewan; lots of fam­i­ly here, so we do this reg­u­lar­ly. Any­how, my Mom in Regi­na us­es dial-up, which mail vol­ume has made im­prac­ti­ca­ble for me. So, I was pret­ty well re­signed to be­ing most­ly Net-free. Ex­cept for, I’m sit­ting in Mom’s liv­ing room here in a nice or­di­nary res­i­den­tial neigh­bor­hood and I turn on the com­put­er to check some­thing and here’s this WiFi net­work named “jamaville,” the sig­nal is on­ly re­al­ly sta­ble in one cor­ner of the liv­ing room, but hey, I’m on-line. I told my Mom the name and she drew a blank on all the neigh­bors and even called one who’s kind of high-tech, no idea who it is. The word doesn’t even show up in Google (that is, it didn’t un­til to­day). So who­ev­er you are, Ja­mav­ille, thanks and my apolo­gies for leech­ing a lit­tle band­width for the next cou­ple of days. Hm­m, if WiFi is show­ing up in ran­dom lo­ca­tions in small Prairie cities, doesn’t that mean it’s pret­ty soon go­ing to be ba­si­cal­ly ev­ery­where? [Up­date: Follow-up on Google prop­a­ga­tion weird­ness and content-replication lame­ness.] ...
 
IETF News · The IETF, de­spite the fact that it keeps the In­ter­net run­ning, doesn’t get much news cov­er­age. I sus­pect that’s part­ly be­cause its struc­ture is pret­ty well 100% opaque and in­com­pre­hen­si­ble to out­siders (and it seems to not a few in­sid­er­s, es­pe­cial­ly new­bies like me). There’s ac­tu­al­ly no such or­ga­ni­za­tion as the IETF but there are (in al­pha­bet­i­cal or­der) CNRI (who run the Sec­re­tari­at) and Foretec and the In­ter­net Ar­chi­tec­ture Board (IAB) and the In­ter­net As­signed Num­bers Author­i­ty (IANA) and the In­ter­net Cor­po­ra­tion for As­signed Names and Num­bers (ICANN) and the In­ter­net Engi­neer­ing Steer­ing Group (IESG) and the In­ter­net Re­search Task Force and the In­ter­net So­ci­ety (ISOC) and the RFC Edi­tor. Any­how, there’s news. Big news, I think. Most of it I don’t un­der­stand and the parts that I do I may have been told in con­fi­dence. The fair­ly se­vere angst com­ing out of this spilled over in­to the Jab­ber chan­nels from Thurs­day night’s ple­nary ses­sion. I think it may be as smooth and sim­ple as the IETF try­ing to do the same thing more ef­fi­cient­ly, or it could be a lot more com­pli­cat­ed and ug­ly. Don’t ask me to ex­plain it; but I think it mat­ter­s.
 
Telephone Inflection · Just a cou­ple of da­ta points. I spent an hour talk­ing to Hen­ry Sto­ry in France this morn­ing, he’s do­ing some work on BlogEd and giv­en the ob­vi­ous pain around Web au­thor­ing, maybe we should be look­ing clos­er at that. Oh, Hen­ry and I didn’t use the tele­phone, that would cost mon­ey, we chat­ted face-to-face across eight time­zones via iChat AV, which is of course free. Other tele­pho­ny news: I have a Von­age phone on my of­fice desk, and now Von­age tells me that for an ad­di­tion­al C$12.50/­mon­th, I can get a SoftPhone. If I do, then my com­put­er will have a phone num­ber... that phrase some­how res­onates; I don’t think I un­der­stand what it means yet. Mean­while, Rus­sell Beat­tie has been telling us about the present and near-future (1, 2, and 3) of wire­less. It seems to me that the whole world-wide tele­phone busi­ness has been smashed in­to lit­tle pieces and thrown up in­to the air, and who can tell what it’s go­ing to look like when it all land­s. [Up­date: Later the same day, the launch of Skype 1.0 for Win­dows. Hold on tight.]
 
Stupid Broken Wintel · Lau­ren is go­ing on a trip in a few days, and we thought that free video­phon­ing would be nice to have while she’s on the road. She’s got a de­cent IBM Thinkpad X31 which has FireWire, and AIM is sup­posed to work with iChat A/V, so this should be pos­si­ble. You can’t plug an iSight in­to Win­dows, so she got an “iBot” FireWire cam from Orange Mi­cro. The Thinkpad has a FireWire port, but it’s a pas­sive four-wire that doesn’t pro­vide pow­er, so she had to get a pow­ered hub (this is al­ready too much work). Then she fired up AIM and... no luck, it didn’t rec­og­nize the iBot; it did, how­ev­er, let us au­dio chat back and forth. Win­dows XP saw the cam­era just fine and you could see the video. We checked the iBot with the Mac, it worked per­fect­ly with iChat. First, she spent some time on the Web Site; no help, but there was a phone num­ber. She called it, pressed the num­ber for tech sup­port and a record­ing said “for tech sup­port send email to support@.” So she sent the email and got back an auto-responder say­ing “for tech sup­port go to the Website.” Then she tried it with Ya­hoo Mes­sen­ger, and it worked in we­b­cam mod­e, in­cred­i­bly slow, but then there was no au­dio chat. Fi­nal­ly we plugged in our big hand-held video­cam and it al­so didn’t get rec­og­nized by AIM. So I’m not sure whether IBM sucks or XP sucks or AIM sucks or the iBot suck­s. But it’s the iBot that’s go­ing back to the store to­mor­row. (Lau­ren would hap­pi­ly get a Pow­erBook ex­cept for she finds a touch­pad mas­sive­ly un­pleas­an­t. Since it is well-known that wom­en have gen­er­al­ly more sen­si­tive fin­ger­tips than men, I wouldn’t be sur­prised if she’s not the on­ly one. Earth to Ap­ple...) Some­times I hate com­put­er­s.
 
iChat AV Irritation · So, I can video-chat to Si­mon Phipps in his place in Southamp­ton, Eng­land just fine. Steve Gill­mor can video-chat to Si­mon too. But Steve and I can’t video-chat to each oth­er. Con­nec­tionDoc­tor says Steve’s not re­spond­ing to my UDP pack­et­s. NatCheck says nothing’s wrong. I turned off my fire­wal­l, but I’m sit­ting be­hind a DSL mo­dem and a D-Link router that may be screw­ing things up. Mark me puz­zled. [Up­date: Flip Rus­sel­l, who’s been around this track, writes to tell me that here and there around the In­ter­net are routers that just don’t route UDP, and that if one of those is be­tween you and who­ev­er you’re try­ing to iChat with, au­dio/video just isn’t go­ing to hap­pen, and there’s prob­a­bly noth­ing you can do about it.]
 
More Spam Damage · I sub­scribe to a few IETF-centric mail­ing list­s, in­clud­ing atom-syntax, that are host­ed at the IMC; Re­cent­ly, I post­ed an Atom-related sug­ges­tion (see al­so here) but the post­ing nev­er showed up. I checked with Paul Hoff­man, who does a fan­tas­tic job of run­ning the site (competently-run mail­ing lists aren’t that easy to come by), and he told me I’d been auto-unsubscribed be­cause my ISP had de­cid­ed, de­fer­ring to the SpamCop Block­ing List, that the IMC’s SMTP was a spam source and (tem­porar­i­ly) black­list­ed it. This is the sec­ond time in a month that I’ve had le­git­i­mate emails get bounced this way. I re-subscribed from my Sun ad­dress, but ap­par­ent­ly, to quote Paul: “If you trust spam­cop, you will con­tin­ue to lose your mail ran­dom­ly with­out know­ing why.” If you read the SpamCop page linked to above, it looks like SpamCop would agree with Paul’s anal­y­sis. Any­how, now I’ll find out if my ISP will stop ap­ply­ing this fil­ter to my email, and if not, I’ll have to find a new ISP.
 
It’s Over · When in­vi­ta­tions to LinkedIn and Orkut start­ed com­ing along, I joined up, and (on Orkut) even proac­tive­ly signed up to be friends with some peo­ple I knew who were al­ready there. But I nev­er fig­ured out what it was all for. Maybe I’m wrong, maybe these are the first symp­toms of some­thing big. To­day, I look at my in­box and see a few more in­vites and you know, it just isn’t worth the time to go hit the sites and click on “Yes” or “No.” So if it looks like I’m ig­nor­ing your in­vi­ta­tion I am; but don’t take it per­son­al­ly.
 
New Virus, Yow · This virus that’s go­ing around try­ing to get you to click on a .PIF be­cause it’s par­tial or has non-ASCII in it or what­ev­er, it’s hit­ting my in­box hard­er than it’s ev­er been hit be­fore. Maybe we’re watch­ing a record be­ing set; the mind bog­gles at the thought of the bil­lions of these things that are swoosh­ing around mailspace. The sto­ries say the virus-hounds haven’t fig­ured out what it does yet. Wel­l, I can tell you one thing it does: mail it­self to ev­ery­one in your ad­dress book, fak­ing a sig­na­ture from some­one else in your ad­dress book. Er, maybe the world should stop us­ing Out­look. Just a sug­ges­tion. [Up­date: Jon Udell writes me to point at this ex­cel­lent four-year-old piece; the email client there is Out­look and he still us­es it, and the way he us­es it makes it safe. He’s right and I was wrong: The prob­lem isn’t Out­look, it’s Win­dows.]
 
Get Yer Social Networking Here · Some­time in De­cem­ber, some­body flipped a big switch and all of a sud­den ev­ery­one was invit­ing me to join their Linkedin net­work. Then sud­den­ly last week the Kozmick Finger point­ed at Orkut, and near as I can tel­l, all the geeks on the plan­et have spent this week­end busi­ly invit­ing each oth­er to be Orkut pal­s. It all seems most­ly harm­less; mind you, I haven’t ac­tu­al­ly got any use out of ei­ther of ’em. For what it’s worth, all the Orku­tians seem to be heavy geek­s, while about half the Linked­in­crowd is VCs and busi­ness­peo­ple. I don’t think it’s gonna change the world, but I’ve been wrong be­fore. To those whose in­vi­ta­tions I’ve de­clined: sor­ry, noth­ing per­son­al, it’s just that I feel I ought to ei­ther have spent some face-to-face time with you or been in some sub­stan­tial on­line in­ter­ac­tion.
 
Email Tech 2004 · In the last few weeks I’ve up­grad­ed my email en­vi­ron­ment here and there, and I thought it would be worth pass­ing on hints in the ar­eas of mail-sending ser­vice and se­cu­ri­ty ...
 
On Postel, Again · Mark Pil­grim has echoed Aaron Swartz’s ear­li­er call for, in gen­er­al, for­giv­ing pars­ing of In­ter­net con­tent and, in par­tic­u­lar, the ap­pli­ca­tion of this “liberal” pol­i­cy to the pars­ing of sub­scrip­tion feed­s, and in par­tic­u­lar par­tic­u­lar, to the pars­ing of the Atom for­mat. Others in­clud­ing Dave Win­er have weighed in on the oth­er side. Parts of this no-exceptions mes­sage are mis­tak­en and mal­formed, but I’ll parse it for­giv­ing­ly and ad­dress some in­ter­est­ing re­lat­ed is­sues ...
 
Telephony R.I.P.? · I have an iSight and a nice new Mac lap­top. I al­so have a beat-up old Mac and a de­cent Canon video­cam that I don’t use that much, not hav­ing (yet) de­vel­oped videographer’s re­flex­es. Any­how, the Canon has firewire out­put, so I plugged that in­to the old Mac and what do you know, it works just fine with iChat AV. So we put the old Mac and the Canon with a lit­tle tri­pod on a desk in a qui­et but wired area up­stairs and it’s a free video­phone to any­where in the world. Res­tat­ing for em­pha­sis: when­ev­er I’m any­where in the world and have an In­ter­net con­nec­tion, I can have a free video­phone call home, that goes on as long as I need to and nobody’s count­ing min­utes or run­ning up a phone bil­l. Let’s see; free tele­phone with video, or pay-for-it tele­phone with no pic­ture. Cost­ly and voice-only, or free with a pic­ture. I think this is what an in­flex­ion point smells like.
 
DynDNS, Dig It · This is just a plug for DynDNS.org, who pro­vide all sorts of use­ful ser­vices re­lat­ed to DNS and mail. Ba­si­cal­ly they take care of a bunch of things that in an ide­al world shouldn’t be nec­es­sary: If you have broad­band but your DNS keeps shift­ing; if you can’t get to your ISP’s mail-sending SMTP ma­chine; and so on. Some of the ser­vices are free, and the rest are ridicu­lous­ly cheap, and it seems to Al­ways Just Work. Check it out.
 
Debian Problem With PPPOE and Static IP · I’m abus­ing my blog with an ap­peal for help; all those who aren’t De­bian ex­perts in a help­ful mood can skip over this one ...
 
Bye-Bye, Modem · Talk­ing around the ta­ble over sushi and beer the oth­er night, we re­al­ized we’ve reached an in­flec­tion point; a lot of us can re­al­ly no longer use dial-up ef­fec­tive­ly on the road. The vol­ume of email (most­ly spam, but even the le­git stuff) has got­ten so high that you just can’t af­ford the time for the mo­dem to wade its way through it. IMAP plus server-side fil­ter­ing helps less than you’d think, be­cause IMAP is such a chat­ty pro­to­col. It was here in Ja­pan, a decade or so ago, that for the first time I was able to di­al up and get my email from on the road. At the time it was a huge quality-of-life win and I re­mem­ber sign­ing off “Emailing ya from Ja­pan, dig it!” Wel­l, I’m blog­ging from Ja­pan, but it’s through a fast smooth Wi­fi con­nec­tion, just open up the lap­top and I’m on the air. Two of the peo­ple in the room here have Von­age box­es and are get­ting lo­cal phone call­s; dig it!
 
Another Whack at Spam · What hap­pened was, I was at a ta­ble with Jere­my Za­wod­ny, Dave Sifry, and Doug Cut­ting, which is prob­a­bly around fifty ag­gre­gate years of big-iron ex­pe­ri­ence. So as usu­al we were bitch­ing about spam, and we had an idea that would shut it down for sure. [Up­dates: Pri­or art and an im­prove­men­t.] ...
 
Spam Snapshot · My spam fil­ter is rea­son­ably well-trained against SoBig now, such that on­ly a dozen or two forged-To bounces a day get through; the price be­ing that I am un­like­ly in fu­ture to no­tice when my own emails bounce. Al­so, the end­less in­ven­tive­ness of the spam mer­chants is alarm­ing, and they man­age to get a few past Mozilla’s Bayesian de­fens­es. Two items of note in this space in re­cent days: First, the ever-inventive Paul Gra­ham sug­gests that fil­ters au­to­mat­i­cal­ly deref­er­ence ev­ery URI in each in­com­ing mes­sage; this is re­al­ly clever be­cause it means that any­one who sends spam is go­ing to in­cur a re­al tan­gi­ble ex­pense, since no­body gets band­width for free. Se­cond, Bill Wein­man pro­pos­es AMTP, for Authen­ti­cat­ed Mail Trans­fer Pro­to­col; a quick read­ing of the draft re­veals no ob­vi­ous fa­tal flaws but I’m no ex­pert in this space. Know­ing who sent ev­ery email would very like­ly make the prob­lem go away, though.
 
WiFi Where? · I was sit­ting in the air­port us­ing my $6.95/­day Way­port con­nec­tion to read the news, which in­clud­ed a sto­ry about the dan­ger of a WiFi bub­ble. Which leads one to won­der, where do we re­al­ly need it? ...
 
Degrees of Viral Separation · I was nuk­ing a few dozen pieces of So­big.F dung that had made it through the mail fil­ter, rec­og­niz­ing a few of the names that had been forged in the “From” and “To” field­s, not rec­og­niz­ing more, and it oc­curred to me that each in­stance of this virus con­tains an as­ser­tion about three peo­ple: that one of them knows the oth­er two. There’s a busi­ness plan lurk­ing in here ...
 
On Email · Every­body knows that this week’s virus-storm has hit so hard be­cause ev­ery­one runs Out­look; so one way to im­prove the sit­u­a­tion is to not run Out­look. Here­with con­sid­er­a­tion of some pros and con­s, and a look at a few email al­ter­na­tives, in­clud­ing Eu­do­ra, Mozil­la, and Pe­ga­sus ...
 
Thinking of Bob · Bob Met­calfe, that is, who reg­u­lar­ly pre­dict­ed that the In­ter­net would col­lapse, and fa­mous­ly ate his words—literally, he ground ’em up in a blender and ate ’em, I was in the room—at one of the Web con­fer­ences. Wel­l, at the mo­ment I can’t con­nect to AIM and I can’t con­nect to Google, and I’m get­ting a cou­ple dozen echospam per hour (bounces from spam with my ad­dress forged) and even though I just turned on the com­put­er af­ter a cou­ple hours down­time, NetNewsWire says “nothing new” which I flat­ly don’t be­lieve. Prob­a­bly just some­thing else flail­ing away at the Win­tel mono­cul­ture. But I don’t like it.
 
author · Dad · software · colophon · rights
Random image, linked to its containing fragment

By .

I am an employee
of Amazon.com, but
the opinions expressed here
are my own, and no other party
necessarily agrees with them.

A full disclosure of my
professional interests is
on the author page.