Mar 30th, 2007 by ravi
In preparation for a day of appreciation

April Foos RFCs
My friend Tom has a new book, this time around in a lighter vein. If you are a geek with some history, you will find it worth a read. Check out his Live Journal post for more info.

Read the full post and comments »
Nov 14th, 2006 by ravi
Old and bloated!

Slashdot has a post today on the Web turning 16, which links to W3C’s history of the web page. Back in 1994 or 1995, shortly after my site was listed in a top 100 list (I think PC Computing Magazine?) some guy contacted me with a bit of interesting info: according to his records, my CERN HTTPD server was among the first 250 sites (I forget how he calculated that — perhaps based on the old Yahoo index?). Reading through the W3C history page I see:


January: Around 50 known HTTP servers.
October: Over 200 known HTTP servers.

I started up my server in early February 1993 and shortly after sent information about it to the WWW mailing lists of the time. Depending on what is meant by “known” above, It seems quite possible that the guy was right about my making the first 250… maybe…

Read the full post and comments »
Oct 30th, 2006 by ravi
The politics of Internationalised Domain Names

Vint Cerf is making noises that IDN is a huge technical challenge:

“One of the most important aspects is for the user to make unambiguous references to every registered domain name.

“Historically this has been through a small subset of Latin characters.”


Mr Cerf said that in order for other scripts to be introduced into the domain name system, there needed to be rigorous testing to ensure that users could be certain they will reach their online destination no matter which script they used.

“Domain names are not general natural language expressions. They are simply identifiers,” he said. “They must be unique. Names registered today must be able to work into their distant future no matter what characters are added.”

He warned: “A miss-step could easily and permanently break the internet into non-interoperable components.”

I respect Cerf but this seems like fear-mongering (perhaps to counter international pressures particularly on ICANN, which is today controlled by the USA) rather than a technical argument. Uniqueness of names can be guaranteed in IDN, and talk of “permanent” break of the Internet into non-interoperable components, is a bit irresponsible. Also, phishing/spoofing attacks (the concern brought up above regarding the certainty of users in accessing sites) are not unique to IDN and have been addressed both before and also within IDN. Wikipedia offers a decent introduction to IDN/IDNA that addresses many of these points, and provides information on IDNA support in applications (e.g: Mozilla/Gecko).

The opinion of Viviane Reding of the EC, quoted in the same article, are, I think, a bit more on target:

Viviane Reding, the EC’s information society commissioner, said: “Bridging the digital divide is not just a matter of screens and cables.

“It is equally important to recognise the extent and value of cultural diversity within global village of the internet. That is why multilingualism is important.”

She said that IDN was “sometimes wrongly seen as technical issue”.

“There is legitimate political imperative,” she said. “Users want to be able to use Chinese ideograms and Arabic scripts.

“There is a real danger that a prolonged delay in the introduction of IDN could lead to fragmentation of the internet name space.”

I cannot but draw parallels to the (oft-mentioned) doomsday protestations of car manufacturers regarding everything from seat belts to better mileage.

Read the full post and comments »
Oct 7th, 2006 by ravi
$11 million judgement against Spamhaus

A bit of slightly old news: A judge in Illionois awarded [default] $11 million to a person/entity listed as a spammer in Spamhaus’ block list. There are many interesting issues here, primary of which is that of jurisdiction. Below is Spamhaus’ response:

The Spamhaus Project – Answer (2) to David Linhardt aka e360 Insight LLC

David Linhardt (aka e360 Insight LLC) filed a lawsuit in an Illinois court with no jurisdiction over the United Kingdom and obtained a default judgement ordering Spamhaus in the United Kingdom to pay Linhardt damages, to remove Linhardt’s ROKSO record and to cease blocking Linhardt’s spam. Unfortunately Mr. Linhardt was not advised that U.S. court default judgments have no validity outside of the U.S.

Below is ArsTechnica‘s report on the affair:

Spamhaus fined $11.7 million; won’t pay a dime

An Illinois judge has ruled that UK blacklist site Spamhaus must pay $11,715,000 to an alleged spammer. The ruling, issued Wednesday, comes after e360insight sued The Spamhaus Project in the Northern District of Illinois, alleging that it had suffered massive harm to its business as a direct result of Spamhaus’ decision to list e360 on a ROKSO anti-spam blacklist.

How did it happen? After all, the judge, Charles Kocoras, is chief judge of the District Court in Northern Illinois and was last month awarded the Chicago Bar Association’s highest honor, the Justice John Paul Stevens Award. This is not a guy who hands out his verdicts like candy.

The answer is that it happened because Spamhaus didn’t bother to reply.


[ Link ]

Read the full post and comments »
Aug 15th, 2006 by ravi
Farber v Cerf on Neutrality

Computer World has a report on the recent Farber and Cerf debate on net neutrality at Center for American Progress. I think Farber’s warning is appropriate:

Farber urged people on both sides of the net neutrality debate to approach the complicated issue with fewer TV ads that aim for emotions over facts. There’s “too many bumper stickers, too much noise, and not enough facts in digestible form being put up there,” he said.

[Link to Audio]

Read the full post and comments »
Jun 29th, 2006 by ravi
Google Checkout


The latest thing from Google: Google Checkout. An e-commerce (what a quaint old term ;-)) payment system. One nice feature I notice is that Google will help anonymize your email address.

Read the full post and comments »
Jun 10th, 2006 by ravi
No Apple for India

BW: India: Why Apple Walked Away
[…] Apple Computer Inc. has shelved plans to build a sprawling technical support center in Bangalore, even as IBM (IBM ) and other tech powers are ramping up. Just three months back, Apple appeared to be on the same trajectory, and there was talk of the company hiring 3,000 workers by 2007 to handle support for Macintosh computers and other Apple gear. Many in India even speculated that Jobs might travel there this year to publicize Apple's commitment to the country.

It wasn't meant to be. In late May, Apple dismissed most of the 30 new hires at its subsidiary in Bangalore. (A handful working in sales and marketing will stay on.) Spokesman Steve Dowling would say only that Apple had "reevaluated our plans" and decided to provide support from other countries. Another source familiar with the situation, though, says the decision was cost-driven. "India isn't as inexpensive as it used to be," the source says. "The turnover is high, and the competition for good people is strong." Apple feels it "can do [such work] more efficiently elsewhere."

The shutdown highlights concerns about the sustainability of India's fast-track economy. True, India grew 9.3% last quarter and is still home to the world's largest and fastest-growing offshore outsourcing sector, which last year generated some $17.3 billion in revenues and employed nearly 700,000 people, according to the McKinsey Global Institute. Yet India's benchmark Sensitive index, or Sensex, has dropped by 20% in the past month as global investors have fled emerging-market stocks. And the outsourcing sector is now plagued by concerns about rising wages. Entry-level pay at tech and outsourcing companies climbed by as much as 13% annually from 2000 to 2004, while salaries for midlevel managers jumped 30% a year during the same period, to a median of $31,131, according to McKinsey and Nasscom, India's software industry association.


Read the full post and comments »
Jun 9th, 2006 by ravi
MPAA gets sued!

Just for the laughs:

TorrentSpy suit accuses MPAA of hacking


TorrentSpy alleges that in July last year the MPAA paid the hacker $15,000 for the information. TorrentSpy also alleges that the MPAA told the hacker it didn't care how he got the information and that it would protect him from any liability in obtaining the information. The suit does not explain how TorrentSpy discovered the information breach.


Read the full post and comments »
May 26th, 2006 by ravi
Net neutrality moves forward in Congress

CW: House panel approves Net neutrality bill
May 25, 2006 (IDG News Service) — A U.S. House of Representatives committee has approved a bill that would prohibit broadband providers from blocking or impairing their customers' access to Web content offered by competitors.


Read the full post and comments »
May 11th, 2006 by ravi
Using and synchronizing contacts

Most of you probably have a desktop computer, perhaps also a laptop, one or more hand-held device (Palm computer, iPod, etc), and a mobile/cell phone (and there is also the home phone, but I will ignore that here). The standard problem: keeping the information sync'ed up between all of them, without needing data re-entry.

What data?

At the least, contact/addressbook information, and calendar/task entries. Stuff that falls under the PIM (Personal Information Manager/Management) cloud.

How is it accessed?

You would think that in the Internet age you would store the information on a central server and access it using standard protocols supported by client applications. That, it turns out in my experience, is tougher than I would have thought.

The technology

If you live in a pure Microsoft world (Windows on your PC, laptop, handheld and mobile phone) you probably can stop reading, at this point, and add a comment exhorting me to come over to your side! Windows probably does a decent job of Sync'ing between your computer and your handheld or mobile phone. Throw in an exchange server and you probably get syncing across computers as well. Well, what about the rest of us?

The standard technologies (well, one of the standards: as the saying goes, the great thing about standards is that there are so many to choose from!) for contact and calendar information are LDAP directories and iCalendar based calendar subscriptions and import/export.

Surprisingly both LDAP and iCalendar are supported by today's addressbook, email and calendar applications, including: Mozilla Thunderbird, Mozilla Sunbird, Apple Addressbook (and hence Mail), Microsoft Outlook (from what I know), the many GNU/Linux/GNOME/KDE applications (Kmail, Evolution, Kontact, etc).

Well, are we done, then?


Unfortunately there are many annoyances to deal with:

  • There aren't many free or commercial LDAP directory services available on the net. In fact, the only one I have found, which I highly recommend, is ScheduleWorld (which provides not just LDAP directories, but also standards based calendar service). This general lack of LDAP services pretty much nixes sharing your addressbook.
  • Mozilla Thunderbird (to my knowledge) does not support lookup across multiple LDAP addressbooks for address completion, making use of LDAP just another bit tougher.
  • Access to remote information is not a viable option today on iPods and various mobile phones. They do not support LDAP or iCalendar access to your information. The good news is that many of them provide two way synchronization (where applicable) using applications and drivers on your computer.
  • There are a large set of iCalendar based calendar services available (e.g: the aforementioned ScheduleWorld, Google Calendar) As always there is a catch… well two in this case:
    • Most calendaring services use a publish/subscribe model and not a synchronization model. In other words, you can publish your Apple iCal or Mozilla Sunbird calendar to the server, or subscribe to your server calendar on one of these applications, but you typically cannot update either willy-nilly and have them synchronize with each other.
    • Free/Busy information: in order to schedule events involving multiple individuals, their free/busy information needs to centrally stored, accessible to the others. In a web-only system, this is trivial. In our multi-tool scenario, free/busy information needs to be synchronized.

Is there hope?

There is hope for the future, but my investigation has found nothing with enough coverage to make it worthwhile. The exception, if any, is ScheduleWorld. As far as I know, these are your options:

  • Use a fixed write (write the information using only one interface), publish to server, and subscribe from many, model: either store your data in an application that supports pushing it to a server or store it on the server. Subscribe from all relevant clients.
  • Use a service like ScheduleWorld which gives you: web-based read/write access to both calendar and contacts, a Java based multi-platform client to accomplish the same, and open interfaces (LDAP, iCalendar) for your multiple client applications. Beware of Apple Addressbook weirdness: not only does it often not import LDIF entries (from LDAP directories), it also does not sync LDAP directory entries to the iPod).
  • Wait for better SyncML support.
  • Hack up your own or use open source scripts to import data into client applications from public services such as Google Calendar.
  • Use a service like Plaxo (which has come under much scrutiny and criticism, all of which you can read easily through a Google search on Plaxo) which synchronizes (only your contact list though) across multiple platforms (Apple Addressbook, Mozilla Thunderbird, Outlook) and provides a web interface.
Read the full post and comments »