Tuesday, November 1, 2005

Pluto has moons?

In a desperate, vain attempt to maintain status as a planet, Pluto has apparently produced some moons: BBC NEWS | Science/Nature | Two new moons found around Pluto (yes, I realize Pluto already had a moon, but its planetoid status has been in question since the discovery of 2003 UB313).

e.p.c. posted this at 20:55 GMT on 1-Nov-2005 .

Thursday, November 3, 2005

Attention Massachusetts Residents: Your State Senate is Selling You Out

The Massachusetts state I/T department decided recently to standardize all documents on something called Open Document Format or ODF, replacing a defacto standard of various Microsoft product formats. Microsoft has, understandably, been upset about this and has pulled out all the stops to stop this from going forward. Totally coincidentally (right), a State Senator added a line to a recent appropriations bill which would effectively kill ODF and possibly standardize solely on Microsoft formats: Bill drops bomb on OpenDocument Format in Massachusetts: Two days after a Senate oversight committee in Massachusetts (1) questioned the authority of the state's IT department (ITD) to standardize on formats for storing public documents and (2) demanded that state officials take more time to study the potential impact of setting the OpenDocument Format (ODF) as a standard, an economic stimulus bill that goes before the Massachusetts Senate tomorrow (Thursday, Nov. 3, 2005) has been suddenly amended with text that, if passed, would essentially subjugate all IT procurements and ITD decisions including standard setting to a special task force.

Why should this concern you?

  • Standardizing on a proprietary data format locks you into the proprietary software that reads that format. While there are many software packages which attempt to read the various MS formats (*.doc, *.xls, etc), they are dependent on reverse engineering the format, MSFT can pull the rug out at any time by updating their software and changing the underlying format.
  • Standardizing on a proprietary format locks you into the proprietary software that reads that format (yes, this is a repeat, but read on): you have to use the software as issued by the company to read your own data. If the company changes the format, you have to upgrade all of your licenses (to keep data formats in sync). If the company goes out of business you are screwed. I was advising a colleague of a friend of mine over the summer who was using a proprietary package to manage his restaurant. All of his menus, recipes, accounts were managed by this product. The product had a built-in die-by date and promptly stopped working at the end of June, 2005 (years after he'd acquired the product). He tried to reach the company to get an updated version and surprise, the company is out of business. There was no follow-on company. There was no warning that he was about to lose everything he needed to run his business, it just stopped working. This was with a legitimate installation of a proprietary package, license paid up, etc. He had no legal, legitimate recourse to recover the data. Had the data been in an open, documented format he could at least try repurposing it in another application. Proprietary document formats tie you to proprietary software and to the whims and fortunes of the software provider.
  • Proprietary software is expensive, even with various bulk licensing regimes. Governments are always under the gun to cut costs and lower taxes, so it is surprising to read of a legislator in favour of higher costs and taxes when there is no real need.
  • Open document formats do not necessarily imply open software, free software, non-proprietary software. There is no reason Microsoft products could not be used with ODF documents except for obstinance and instransigence on MSFT's part (ODF is allegedly XML, MSFT products allegedly can read, parse, and save in XML formats).

So, if you live in Massachusetts, contact your state legislators and express your concern about this shift and ask why they are willing to increase the costs of information technology for the state by sticking to proprietary data formats.

e.p.c. posted this at 14:46 GMT on 3-Nov-2005 .

Saturday, November 5, 2005

New competitor to VCs: Google

I added a link to Paul Graham's essay The Venture Capital Squeeze in my earlier post Here's a surprise: Web 2.0 Entrepreneurs don't want VC money, but wanted to highlight some insights he makes:

  1. Starting a startup is significantly cheaper today than ten years ago. Software is cheaper if not free, systems are cheaper, software development is cheaper and easier. starting a startup is so cheap, venture capitalists now often want to give startups more money than the startups want to take.
  2. Sarbanes-Oxley, intended to prevent future Enrons, has set the bar so high for public companies and increased the expense of compliance, that the exit strategy of the IPO (going public) has been pretty much eliminated for the typical startup, making the only viable exit strategy a buyout by a larger company.
  3. Due to #2, the companies doing the acquistitions (Google, Yahoo!, Microsoft) are bypassing the VCs and approaching and interacting with startups directly. The VCs are, in effect, competing with Google for attention and investment opportunities with startups.

I'd add, since it is cheaper to start up a startup, there's less personal investment involved both in time and money. In turn, I wonder if there's less of a personal need for a gigantic return. The risk of going the VC route is in essence higher than the reward of selling out. The whole sell-out as exit strategy meme was mocked at the Web 2.0 conference by various VC and financial industry people, but the reason someone sets out to start up a business is not to ensure a stunning return for a venture fund. It's to solve a problem, have fun and make some money along the way.

§

A followup: VCs vs. The Platforms: Levene mentioned to me later that Google has even set up a fund to compete with VCs for early stage company financing (I had not head this before), and that Yahoo feels it can and must compete to buy early stage companies before VCs can get in with larger financing.

e.p.c. posted this at 20:18 GMT on 5-Nov-2005 .

Thursday, November 10, 2005

AMT: Bad for you, bad for me

Interesting article at CNN Money: Why middle America is getting hit with a tax for the rich - Nov. 10, 2005, stunned me with this quote:

To give you a sense of just who might get caught, this year only 1.8 percent of married couples with two kids and an adjusted gross income between $75,000 and $100,000 will be subject to AMT. Next year, that number jumps to 73.4 percent.

Now, we've been hit with the AMT for the last couple of years (add living in New York to some extraordinary income situations like cashing in a parent's 457 and pension). The most immediate effect has been that our deduction for mortgage interest is mostly wiped out. There's some there, but not much.

Given that one party has controlled the legislative and executive branches for the past five years, you'd think that perhaps, just maybe, they'd take some actual action on eliminating or reducing this problem which will hit large numbers of their constituents in an election year. But they haven't.

e.p.c. posted this at 16:08 GMT on 10-Nov-2005 .

Friday, November 11, 2005

Dear web: Please make your type larger

I'm in my late 30s now. I have been using computers (on and off) since 1978, from an Apple 2 at a summer camp, to a MicroVideo Interact at home, and an Apple 2e which I used to pull myself up to a C in Physics at my high school.

I progressed to PCs in college and a Mac later, using PS/2s at IBM. In all, just under twenty years using CRTs for display. Some time in the late 1990s I switched to a laptop as my primary computer and am now on my third personal Thinkpad (I lost count of how many I raced through at IBM, I averaged one every ten months).

Laptops don't use CRTs, they use LCDs. The key difference being that instead of having electrons beamed at your eyes at lightspeed from a ray gun, instead they're emitted from flourescent lights behind the LCD screen (the LCD acts as a filter).

Now, I'm not an optometrist, and I concede that this may just be the effect of, ergh, getting old, but I'm convinced that the constant flicker of the LCD is causing my eyes to wig out.

Wig out being a technical term for: I find it increasingly difficult to read many web pages that tweak the font size for body copy down to 10 points. I'm constantly hitting Control-+ (in Firefox) to kick up the font size a couple of points, and this is with my reading glasses on.

Now, I go to the optometrist every year and every year he insists that my eyes are fine. My distance prescription has barely budged in ten years, my reading prescription degrades a little bit but I can still use my glasses from five years ago (if it were not for Frisket having used them as a chew toy).

Maybe it's a factor of the resolution of the LCD, maybe it's due to the flicker of the lights behind the LCD, maybe it's my eyes or age, I don't know. I do know that I'm justthisclose to using a user-stylesheet to override the font sizes even though that will likely screw up the layout that someone has labored over for days.

e.p.c. posted this at 00:39 GMT on 11-Nov-2005 . , Comments [1]

Sunday, November 13, 2005

NY1 needs a copy editor

The NY1 web site has been useless for the past several days, unresponsive to various browsers or even a simple HTTP/1.0 check. This afternoon I finally load it and see: Screenshot of ny1.com taken 20051113T1919Z.

It's Pay to Pray, though given recent problems with various faiths' ministers, perhaps it wasn't a mistaken spelling.

e.p.c. posted this at 19:21 GMT on 13-Nov-2005 .

Thursday, November 17, 2005

Hit & Run for Thursday 17/11/2005

Heading to Boston for a friend's departing-the-blue-cloud party.

I don't know what to make of the new ZFS filesystem from Sun for Solaris/ OpenSolaris. Here's a flash tutorial. Via ongoing.

ZFS presents a pooled storage model that completely eliminates the concept of volumes and the associated problems of partitions, provisioning, wasted bandwidth and stranded storage. Thousands of filesystems can draw from a common storage pool, each one consuming only as much space as it actually needs. The combined I/O bandwidth of all devices in the pool is available to all filesystems at all times.

I guess my most immediate question is how open is the OpenSolaris license?

Ate dinner with some friends last night. In between recalling days in Baker 168 at CMU, they recommended a photo site they've been using: Phanfare. I'm a flickr fan (phan?) myself, but Phanfare apparently also supports video, so I plan to look into it over the next couple of days.

Top on my to_read list at del.icio.us:

My current projects are: dripldu: sort of social notification meets frequent flyers meets social networking. Not a dating service. YaSoBoTo: Yet Another Social Bookmarking Tool: sort of back-burnered for now, but basically bookmarks + ratings + site statistics, all rolled up nicely for people looking for marketing data. Spotmarket.tv: marketplace for spot advertising also somewhat backburnered given Google's entering-not-entering the space. Basically a marketplace for spot advertising on cable systems.

I'm also winding down a couple consulting gigs.

e.p.c. posted this at 23:38 GMT on 17-Nov-2005 .

Friday, November 18, 2005

!@&$&@# iPod

So, just after writing the previous post, I undocked my iPod from the Mac and proceeded to walk out of my office with it and some other things.

I placed it in my black computer bag.

A few minutes later, I stuck my cellphone into the same bag, and noticed an eery white light. The iPod was all lit up. Huh I thought...wasn't it sleeping and locked? Hmmm

I tried turning it off (well, to sleep) and no luck.

So, I replaced it in the bag and carted all of my stuff to the car for the drive to Cambridge.

Somewhere around Stamford, CT I pulled over for gas and decided to look again at the iPod. Sure enough, it was still all lit up and stuff, no disk sound, nothing going on, no response to the various control-wheel options I've used before to kick start an iPod.

On arrival at the hotel in Cambridge, I pulled it out again and, oh joy, it wasn't on anymore. So I tried to turn it on.

The black apple comes up, there's some clicking sounds, and then it shuts off. Clicking is not a sound I want to hear from any hard drive, let alone my iPod.

I will borrow Alister's charger in the morning and see if it's just a battery drained issue, but the clicking is a troubling sound that hasn't occurred in the past when the battery was merely drained.

argh. I say, argh.

Luckily I brought the Sirius Sportster I'd bought earlier this year and spent the drive listening to a mix of 80s tracks and trance/dance/techno.

e.p.c. posted this at 06:23 GMT on 18-Nov-2005 .

Sunday, November 20, 2005

Little shivs want words, pushed it right back

Woke up this morning to a bright, clear day in Cambridge. Took Frisket out for a long rambling walk along the river into Kendall Square and around the neighborhood east of Kendall. Today's earworm was about three lines from Harborcoat by R.E.M.

I made the mistake, just before leaving, but with Harborcoat sort of rambling in my head, of looking up the lyrics online.

I'm not sure why, just now, I thought to look them up, after all I've known them for years.

I don't know if they're the right lyrics of course, but in R.E.M.'s early days the lyrics were more like additional instruments to the music, not something separate.

Perhaps that's why you don't hear so many covers of R.E.M.

In college, my friend John and I tried to work out the lyrics to various R.E.M. tracks, initially by actual postal mail, but at some point in either our Junior or Senior years we managed to figure out how to email each other.

Now, this was 1987 or 1988. One didn't just email someone. You had to know the path to that person. If you were lucky, they were on an ARPAnet system. Then, all you had to do was find a path from your system (unless you were on ARPAnet as well) to their system. If you were somewhat less lucky, you were both on semi-well-known UUCP systems. Then you could try bouncing the mail through uunet.

So, I'd send mail from costello!sir-alan out into the wilderness, routing via ncoast and uunet, eventually reaching John's VMS system at Marquette.

Anyway...for awhile we tried to figure out the lyrics to tracks like Harborcoat, and Stumble. By the time we got to Maps and Legends and Cuyahoga, the lyrics were becoming more distinct, less part of the music and more a specific thing. I still liked R.E.M. as they went mainstream, but their music definitely changed. Not for the worse, just different.

Anyway, Frisket and I set out on a little ramble along the river. We're staying at the Hotel Marlowe which is on the east side of the Cambridgeside Galleria, across from the Sonesta and the old Lotus Development buildings.

An aside on the Marlowe... when we first stayed here I described it as a W Hotel gone tragically wrong. As though someone stayed at a W, and then through a very bad phone connection described it to someone else, who then transcribed it in a language they only partially understood, and then sent it off to a designer to implement.

Most of my original criticisms still stand. But we've adapted. They truly, honestly accept Frisket as a guest (though this trip they didn't add her name to the welcome list by the biscuit bowl at the front). And they have honest-to-G-d free broadband. The only improvement I'd make to the broadband is to eliminate the accept-our-terms page you have to go to each time you connect to the network. Oh, and they have 802.11g throughout the hotel, so you don't even need to use the in-room ethernet.

The room-service options are still not what I'd expect at a hotel of this caliber elsewhere. The food is fine...just the options are limited, especially if you arrive after 10:00 p.m. And the decor is still rife with a strange animal-fur-leopard-spot motif.

But we like it, it's close to our friends, and it used to be a good spot for Lisa when she was with IBM.

So (this is a rambling post by the way, punt now or accept it), we head off down to Canal Park along the river.

We walked the length of the park, running into a couple of dogs along the way.

We then cut across to Broadway and walked up Broadway to a walkway to 6th street.

We took 6th to Charles St. where Frisket got to talk to several other dogs. I'd thought it was a dog run, but it was a soccer field which just happened to have a number of dogs and their people, watching a soccer game get set up.

We headed back along Hurley Street and finally found the squirrel Frisket had obviously been tracking all along. The squirrel was hanging out at the playground. I may have let Frisket run around a bit chasing the squirrel, of course she at no time entered the playground, which is off-limits to dogs, even when empty.

For the remainder of the walk Frisket was quite pleased with herself, and continued to look for Squirrel the remainder of the way down Hurley to First Street.

The Best Buy at the Cambridgeside Galleria Mall either suffered an automotive intrustion overnight, or just felt a strong need to replace their doors (perhaps in preparation for crowds on Black Friday?)

On returning to the hotel Frisket was offered many treats, the desk staff at this point know her by name (though we're just Frisket's People, I wonder if we're registered as that. Maybe next checkin we just let Frisket leave a paw mark on the registration book.)

e.p.c. posted this at 04:10 GMT on 20-Nov-2005 .

Sunday, November 27, 2005

Ten Privacy Protection Tips from EPIC

From the Electronic Privacy Information Center, Hoofnagle's Consumer Privacy Top 10 an excellent list of steps to take to safeguard your privacy:

  1. Opt out of prescreened offers
  2. Stop your phone records from being sold
  3. Keep your banking records private
  4. Get free credit monitoring
  5. Add your phone numbers to a do-not-call registry
  6. Safeguard your social security number
  7. End student profiling
  8. Avoid loyalty programs
  9. Secure your accounts
  10. Engage in privacy self-defense

(read the article for the complete how-to details)

I've dramatically reduced the amount of junk mail I received by using one of the opt-out services (I thought it was through Public Citizen but can't find any links there). I regularly check my credit report (I actually pay for a service, which I realize is silly now that you can get the reports for free). My interest stems from some early research I did into databases and privacy (circa 1991) and from the discovery that my credit report had bits of my father's credit history included (there apparently was no sanity check when adding my 1972 charge card for Polk Bros. to my account even though I would have been about five).

People believe that privacy is an issue solely because of the Internet, but it's really because of the growth in technology and power of computers that occurred in parallel with the Internet. It used to be hard to create a complex database of people and various bits of identity information. Now I can cart around (and did briefly over the summer) a voter list of one of the area cities for one of the campaigns I worked with. That database managed to contain name, address, phone, birth date, voting record, and whether the person had moved recently.

All of that was public information, and has been for years.

What has changed is that now an individual can create and manage the sort of complex database that was stretching mainframe computing resources just ten years ago.

What has changed with the information is that, while public, the intersections, or in database speak, the JOINs of the various bits of public information create new bits of information. In the run-down from the recent mayoral election, comments were made about the size of Bloomberg's voter-id database. But, given enough money and resource, anyone can build a comparable database. The information is public, though not necessarily free, and the more information you gather the more complex the database is, but for a small region (Say a town or county) it's quite possible to build and maintain a voter list for several thousand dollars (and once initially built the ongoing cost is maintenance).

The value isn't the individual record of information, it's the inferences you can make by combining information.

Take the census data for an area (which is broken down to census "blocks" which may or may not map to a city block). You get all sorts of information from income levels, education, to race and gender identity. Mash that against your list of people (whether a voter list or a list of customers, or a list of people who've moved into the area) and you can make some inferences, some of which will certainly be wrong, but many of which will be right, and you can market against those inferences.

What you are marketing doesn't matter, it could be goods or services, or political candidates, or public issues. Depending on your motives you may want to target people who've moved in (or out) for these products and services, or for ill-intentions.

The Internet didn't bring this issue to us, though it certainly makes certain database activities easier. There is a cost to all of this technology, a cost to the very personalization that we take advantage of on web sites and various frequent customer programs. That cost wasn't clear, though it should have been since any, any advantage an organization can take of the information it gathers, it will take, and it will gather more information to increase the value and the edge.

I see two approaches one could take: be intentionally sloppy about your personal information, so that there's no value in stealing your identity, or be brutally paranoid about the information that you allow to be gathered and maintained about you. I personally tend to the paranoid side.

The only time I give 100% accurate information online occurs when I'm completing a financial or legal task. There is no reason for the pet supply store to know the date I was born. None. There is no legitimate reason for 99% of the websites that request my detailed personal information to have that information. I give out the zip code of my PO box (which is "near" the house), the phone number of our fax line. My birthdate is either 1/1/1970 (shout out to the UNIX epoch) or a day which is close to but not my actual birth date.

The information isn't necessarily wrong, it will all validate, but it isn't correct either.

The important thing is that if someone steals the information from the Pet Supply store it has little value to them. They may get my name and address, which is public in a number of places, but not my birth date (which could be used to access my financial records elsewhere). The important thing is to make the value of the intersections, the data that lives in the interstices, the JOINs, less valuable, less accurate.

Everything one would need to steal a person's financial identity and legel-system identity is already public. My birthdate is available, various home addresses, my parent's names (and mother's maiden name), all are already available. What technology does is to make it possible to aggregate tens of thousands of records of information cheaply. Which would you do, walk garbage can to garbage can collecting information, or find out who in the Pet Supply Store, Bank, or other institution, who maintains the database of customers and attack that weak point?

So, go to the EPIC site and read through that list (apologies to the non-US readers, perhaps an international version of the list should be made). Also consider contacting your US Senator about the Personal Data Privacy And Security Act Of 2005 (the link is to a press release, not sure where the text of the law is). One step I would add to the proposed law is to require anyone maintaining a database of personally identifiable information to allow people to know what information is maintained (by this I mean the types of information, not the actual values) and to allow people to request that the information be purged. I don't think you can allow people to correct information, because how do you certifiably guarantee that the person making the request is the person you have in the database?

One other assignment: if you're remotely into science-fiction, I always recommend John Brunner's The Shockwave Rider for a 1970's dystopian take on a future where everyone's identity is whatever is in the database at the moment.

e.p.c. posted this at 16:53 GMT on 27-Nov-2005 .

Slightly acerbic and eccentric dog walker who masquerades as a web developer and occasional CTO.

Spent five years running the technology side of the circus known as www.ibm.com.

More about me here.

Archives