Traffic Jams

<snip>
Have you ever done the courteous thing when crossing an intersection, and waited back at your stop line because the lane opposite was backed up?  Have you ever had someone making a right turn slip in while you were going out of your way to not block the intersection?
<snip>
As Silicon Valley grew, people started moving into cities farther afield (indeed, they used to be fields), and commute in.  The local city into which they move charges developers for extension of local infrastructure like water, sewer, roads, but there is no model to share that revenue with cities through which the new residents will drive to get to work. So the roads in the middle aren't expanded, and commute times for those who live nearer the jobs go as distant commuters clog the highways.
<snip>
If I could construct cars for free, gas was free, and I had an infinitely large parking lot at my house, I could push cars out my driveway in a steady stream. Where would they go? And why should I not to this? After all, I've paid real estate taxes for my driveway.
</snip>

There's been a fair amount of tearing-out-of-hair lately over the desire of BellSouth et al to create a throttling mechanism for the internet, in essence, to charge for bandwidth based on quality of service (QoS) metrics.

I frankly think that tiering only makes sense.

First, it closely models the real world.  In the real world, I can choose between a free road which may be congested, and a turnpike which may get me to my desitnation faster.  There are toll booths on the turnpike to both encourage the quality difference, and to charge me for the privilege.  Similarly, there are car pool lanes, and fastpass lanes, all rewarding drivers who contribute some (arguable) value to the system with shorter wait times.

Second, much more than in the real world. traffic generation is very very inexpensive in the electronic. If all the T1s to all the content providers were saturated, the trunks would need to be much  much much bigger, as would the switches that route between the trunks and access lines. Some might argue that the content generators already pay for their content to get onto the highway, but like me with my driveway, the physical limits of the driveway permit far more traffic than the system can accept in aggregate.

Now we might argue that this implies a massive increase of scale is necessary for the central infrastructure of the net. Probably so.  But until this the scale is increased there will be those who take advantage of the current system and inject more traffic than their share, like the friendly gent who cut me off by making a right turn on red yesterday while I was waiting for traffic to clear.  And the folks who run the roads will try to find ways (like "no turn on red" signs) to penalize that gent.

In the physical world, the different QoS metrics that exist for physical delivery of envelopes has led to a two-tier system of the government post office and private couriers.  In fact it is multi-tier, as the couriers and post office eash have multiple levels of service at different price points.  So my letter than absolutely must be received by tomorrow, will be received by tomorrow, for a premium carriage fee. Any amount of bulk mail can be pushed into the system by advertisers and not impact my letter at all.

What is so wrong about a tiered approach to net traffic?  Don't we want content we've purchased to arrive faster than spam?

401Ks and other registered plans are not always the best retirement savings vehicles

Disclaimer: I am neither an economist, nor a registered financial planner.  This is not investment advice.

The 401K and other registered vehicles operate under the assumption that future tax rates will be much lower than current. For many, especially comparing income while working to that during retirement, this may be a reasonable assumption.

But consider relatively younger or middle-aged folks with long investment horizons. With a very large federal deficit, underinvested social welfare, and baby boomers leaving the workforce, we seem aimed at a time in which overall tax rates must increase, possibly a lot, in order for the government to continue to deliver social services. If my retirement horizon were 5 years or so this will probably not make much of a difference. In that it is around 20 years, I’m not so comfortable that my marginal tax rate will actually decrease in retirement.

Another factor is deductible mortgage interest. Living on the San Francisco peninsula I have a rather high mortgage (for a modest house, thank you). Itemizing my deductions against income and working through my return, my marginal tax rate drops pretty dramatically despite having a pretty decent income. In retirement I hope I won’t have this high mortgage, but that really means that I won’t have artifically depressed tax rates either.

Add it up. The assumption underlying registered plans such as a 401K are not guarateed to hold. In fact, I'd say they are unlikely to hold for many who are currently working.

In fact, were the US federal government to show any signs of thoughfulness, I might imagine they designed 401Ks with these issues firmly in mind, and that they consider 401Ks an annuity they've purchased in today's dollars to help fund social services in the future.

It may still make sense to invest in a 401K, especially if your employer matches your contributions. But I’ve been directing a good portion of my long-term savings to non-registered plans. YMMV

[Comments from my previous blog]

1. Ralph Galantine left...

Monday, 30 October 2006 8:55 pm :: http://blog.quantmarkresearch.com

The usual advice now is to max out your Roth IRA contribution (contribute after tax dollars, returns and withdrawals are tax free) before you make any contribution to your 401K that isn't matched by your employer. You should also consider maxing out other tax favored accounts, like the education accounts, before contributing to 401K's or traditional IRA's. Regular IRA's and 401K's suffer from the tax issues you mentioned.
Roth IRA's are a good deal unless your income is too high to contribute. Also, if you have a low income year, you can convert some of your traditional IRA to a Roth IRA. The amount you convert is "income" so you have to pay taxes on it, but I don't think there is any other penalty although you will hit an income limit. Your retirement advisor can give details.
On a couple of your other points:
-In retirement, you may not have a big mortgage interest deduction but if you own your home clear you get your "rent" tax free. It is a benefit or "income" to you but it doesn't show up as monetary income so it doesn't get taxed.
-The idea that the 401K's could have been created to generate a tax annuity from the retired living off their pre-tax savings and interest is amusing and might actually help tax revenues as boomers retire. However, it would be a mistake to design that kind of program. 401K holders are also voters and clever people. Account holders may succeed in sheltering their accounts through various devices.

2. Jonathan Bruce left...

Wednesday, 6 December 2006 12:48 pm :: http://jonathanbruceconnects.com/jonatha

So what do mean by non-registered plans? I'd be interested to hear your thoughts on this...

Options backdating is not always unreasonable

Stock options as a component of employee compensation are intended to serve a simple purpose: tie the employee's future to the value of the company, so he is rewarded for any increase in value that he contributes to. Motivated thus, an employee frequently chooses behaviors favouring company performance. "Do the right thing."

The challenge is, how do we figure how much the company increased in value? For that matter, what is a company worth?

The easy measure is the current capitalisation of the company. To wit: multiply stock price by the number of outstanding shares. Very simple.

And being simple, arguably wrong, or perhaps more accurately, sometimes wrong. Notably in cases of speculation.   Create a bunch of buzz around future products and the stock price can run up a lot. Stock prices are benchmarked in terms of company earnings, but Price/Earnings (P/E) ratios can vary wildly depending on speculation on future company growth.  And this is only speculation - the company has no new products or technology (yet), no new revenue (yet), and so is at least arguably not worth any more than it was yesterday or last month.

So in compensating an employee to help build the products the industry is speculating on, what is reasonable as the basis for a grant? To value the company for purpose of that grant in terms of the inflated stock price, where the growth in value driven by the new product is already factored? Or the pre-speculation price so he can benefit from the value of the product he builds?  The pre-speculation price seems to me to be more consistent with the rationale behind granting options in the first place. So one might reasonably back-date the options grant (or perhaps more precisely, the grant price) to the stock price several weeks or months ago, when rumour and news started driving the price up.

None of this is to say that there hasn't, isn't and won't be abuse and fraud in the granting of stock options, but that is being covered pretty thoroughly elsewhere. Use your favourite search engine and you'll find as much as you care to read.

Disclaimer: Working in the tech industry, stock options have been a feature of my comp plans for years. None, however, have ever been backdated.

Manuel, the put-upon Porter

This last weekend was the 12th annual Brew-Ha-Ha and All Round Good Time, a homebrew competition with ballon origami that is held yearly in a friend's back yard, and this year I won for the second year running. The winning beer was a Porter I affectionately named Manuel, after the British comedy Fawlty Towers (though scripts refer to Manuel as a waiter, and arguably he dresses as one, he seems equally a Porter in function)(and, Waiter not being a style of beer, the joke (and likely the beer) would fall rather flat were I to change the name of the beer to reflect literary reality).

I think I only won because of the Chicago Rules of Voting that applied, which have been described as:

• vote early
• vote often
• dead people's votes count twice for sheer tenacity to drag themselves to the poll
• underage voters can vote as well, though presumably they didn't taste. Knowing something about the subject is no more a requirement for our contest than it is for other elections in our experience

This follows because despite winning with my Porter, the weight of the kegs at the end of the day showed that my Pale was actually much more popular, at least in terms of consumption. Go figure.

I'm going to try for 3 years running next year, with a new ale called "Constitutional Challenge", because of course it will take a challenge for me to hold office for three terms.  I'm open to suggestions on what style such a politically important beer should be.

[Comments from my previous blog]

1. Jonathan Bruce left...

Monday, 24 October 2005 12:08 pm :: http://blogs.datadirect.com/jonathan_bruce

Nice work Glen. Are you planning on shipping your brew to the east side?

Is the budget for Shriver's public appearances a legitimate state expense?

I read an article this morning that covers the expense of keeping the first lady of California, Maria Schriver, in speaking opportunities - roughly $500,000 per year. This is reportedly some $100,000 (60 - 180,000 is reported) more than the similar item for the previous first lady.

The defense for this growth in expense?

"To compare Maria Shriver with Sharon Davis," Carbaugh said, is like comparing "Madonna to Jewel. Both are talented women in their own right, but certainly not comparable in terms of notoriety or the sheer number of fans and admirers.''

Well, ok, but notoriety does not constitute a legitimate purpose for the state, or its voters. No doubt Schriver's popularity would create any number of invitations even if she were not the wife of the State Governor. One should instead be asking, and defending, the value of these appearances to the state voters.

I had a look at Schriver's web page, and certainly it includes mention of any number of appearances that have all the hallmarks of worthwhile philanthropy and public service, so Schriver's team has plenty to work with. Though it seems to me that philanthropy is supposed to be a donation - you're not supposed to be compensated for it.

The cynic in me has to wonder when some crafty spinmeister finds a way to defend an unmarried candidate with "s/he'll be cheaper for the state to maintain".

Not that her hubby, Schwarzenegger, is any cheaper ... the report a couple of weeks ago was his events cost $10 - 20,000 a pop, or $622,000 over six months. But even though he's in office, Arnold's expenses are covered by political committees.

Apples to apples: Do comparisons always need to be fair?

I'm becoming increasingly ambivalent about the PC hardware review sites - Tom's Hardware is perhaps one of the better known, but there are a whole bunch of them. This is no doubt partly to do with my changing interests - I'm not quite so involved in hardware as I once was.

But the bigger issue seems to me that the review sites have lost touch with what I do with computers.

Case in point: PC motherboard reviews. I've been playing around with Personal Video Recorder projects for the past 12 or 18 months. In essence, turning a PC into a Tivo workalike. Lots of fun. My latest setup has a big box in a closet with a couple of TV tuners and a big RAID, and a little bitty box beside the TV that has almost no moving parts (there's still one fan I can't seem to shake while still getting decent video output).

It isn't without bugs - but that's what makes it a project.

But one of those bugs seemed, after googling, reading, chatting on email lists and so forth, to be due to DMA issues between a tuner card and the other parts of the system when running large data volume that is generated by a card. in essence, the motherboard or chipset couldn't handle large-volume DMA from multiple sources at once - it caused a reset.

And this got me thinking: if I wanted to pick a great motherboard for home video servers, do the benchmark sites give me any useful information?

The reviews seem to follow a tried and true (and old) formula. Picture of the board. Oh gee, it's green. Isn't that cool. The box is very nice. Look, it has the same slot connectors as everyone. Yup there's room for a large processor heatsink. And so on. Then some benchmarks, which compare this board to others in a raft of performance-critical uses, like spreadsheets, word processing. Okay okay, they do some games as well - very important, the game tests.

But you'll note, no tests at all in which large amounts of bus traffic are driven from an expansion card - all the tests of the system are application-related. And these days, when we're seeing a convergence of home computer and media, application-related use is not the only thing happening.

Another case: I saw a great article - I hate to call it a review - of a couple of desktop PC motherboards designed to take laptop CPUs. The article mentions that these boards are targeting a segment of the market intersted in noise (well, quiet) and power consumption (less of it, that is). Folks like me - I was charmed. "Do processors need to consume over 100 watts?" the author wondered.

Pity was, the article didn't compare any of the boards on the basis of power consumption or noise. The article didn't even report the power consumption or noise of these products. Instead, it included the same old performance comparison.

I think the reason the review folks do this is so they can compare the new product against those they reviewed last week, month, or year. And compare tomorrow's review against this one.

This is nonsensical. In their drive to compare apples to apples, they fail to give the folks for whom these products were built any useful way to select between them.

PS. For those as may be interested, the video server project in question combines MythTV and tuner cards including pcHDTV 3000, AirStar HD5000, and Hauppauge WinTV-PVR-250. The server runs Debian linux and includes a 1.2 TB RAID5 array. The client is a diskless EPIA M10000 based mini-itx system in a Morex 3677 case that boots across the LAN using pxeboot.

[Comments from my previous blog]

1. Sean Kennedy left...

Saturday, 3 December 2005 7:21 am :: http://www.simplemusings.com

Very good observations. I would be interested to know more about your client setup. Can it display HDTV transport streams? Does it output 5.1 sound?

Product management failures: (almost) great coffee

Here's a story of product management gone awry, taken from one of my more self-indulgent pleasures, coffee.

After we recently replaced our coffee machine (ahem), we also bought a new coffee grinder. Our new Krups Burr Grinder does a really good job, much better than any of our previous grinders. The grind is more consistent, and this has to be the quietest grinder I've ever heard (if that isn't an oxymoron). All in all, I love this grinder.

Not to say it doesn't have its faults, and one glaring fault is that the amount of coffee it grinds varies from time to time. It isn't hard to see why, and that 'why' looks to be a textbook case in failure of market segmentation.
The Krups grinder has a little wheel that you use to specify how many cups you'll be grinding, which sets the grind time. This calculation depends on knowing how many beans it will grind through per second. No doubt it works really well on inexpensive coffees, that is, those that come in cans, where the beans are dry on the outside and so slide steadily into the burr.

Better beans, though, or to be more precise, well-roasted beans that are fairly fresh, have oils on the outside. Those oils can make the beans a little tacky, and in this grinder (and others), this tackiness can mean that the beans hang up a bit and the burrs run idle every now and then. A timing-based measure of how much coffee it has ground thus varies, depending on how sticky the beans were that day and on accidents of bean position and so forth. Longwinded explanation, but you can hear it when it happens.

This would have been easy to solve, though no doubt it would have added cost to the grinder. Some rotating vanes or something in the bean hopper would have removed this tendency of the beans to stick, and they would have flowed more steadily into the burr. The ginder even has the vanes, and I thought from the picture that this is what they'd do, but no, they're decoration.

I can see the segmentation error that led to this. The target market for home burr grinders is a sort of mid-market-and-up coffee drinker. But they forgot this target market when deciding whether the hopper design was adequate. 90% of the beans on the general market will work fine in this grinder. But the beans this target market will use won't work as well.

Requirements definition isn't about taking the top priorities from each of a number of different people. I'm reminded of a very old Dilbert cartoon, sales asking engineering for a 23" monitor that fits in a shirt pocket. Requirements definition is about solving most or all the needs of someone - ideally a someone who has a lot of very similar friends. Define a target customer, and delight all folks like that. Take away all their objections to buying. Don't leave them with anything left to think about, or to weigh between your product and someone else's.

Or to put another way, product management is about markets, not customers.

Coming back to the grinder, I do love it. It is head and shoulders above any other grinder I've used, and the only other grinder on the market today that even tempts me costs well over $200 (and it doesn't tempt me that much).