How many engineers does it take to arrange a group photo? All of them. First everyone suggests all possible ideas for lighting, arrangement of people, location. Then everyone argues about the order in which to try all specified permutations, then the camera is accidentally set on 'auto' and takes 40 pictures of everyone arguing. 45 minutes later we had 100 photos of the team, the probability of having taken a good photo finally approaching 1.
Sunday, December 2, 2007
Grant got the memo
How many engineers does it take to arrange a group photo? All of them. First everyone suggests all possible ideas for lighting, arrangement of people, location. Then everyone argues about the order in which to try all specified permutations, then the camera is accidentally set on 'auto' and takes 40 pictures of everyone arguing. 45 minutes later we had 100 photos of the team, the probability of having taken a good photo finally approaching 1.
Sunday, November 25, 2007
the beach in late november
Happy Rocksgiving
While millions of Americans braved the ever deteriorating airlines to visit loved ones, and hundreds of Davisites headed east to Lake Tahoe, this year I was lazy and stayed home. We made a huge thanksgiving dinner, and then rocked out in the garage.
This was the first year I had ever attempted to make a turkey. Last year we celebrated with friends who made a mountain of steamed dumplings, and the year before with friends who welded together a roasting spit from an old motor and bicycle chain, then roasted an entire lamb ("lambsgiving").
Roasting a turkey for the first time worked out well, thanks to the internet. We looked up turkey roasting guides and videos on youtube and managed to average together the sometimes conflicting information into a coherent recipe. We also made potatoes, greenbeans, stuffing, and Jason brought over salads.
So we put the turkey in the oven, and then headed to the garage for some music. It turns out that our average extended jam session length is about the same amount of time you're supposed to wait between basting the turkey. Skip was on the drums, I was on the vibes, and Jason was on the bass, achieving a nice balance. We wrote a few songs, ate a massive amount of food, and then passed out. Good times.
This was the first year I had ever attempted to make a turkey. Last year we celebrated with friends who made a mountain of steamed dumplings, and the year before with friends who welded together a roasting spit from an old motor and bicycle chain, then roasted an entire lamb ("lambsgiving").
Roasting a turkey for the first time worked out well, thanks to the internet. We looked up turkey roasting guides and videos on youtube and managed to average together the sometimes conflicting information into a coherent recipe. We also made potatoes, greenbeans, stuffing, and Jason brought over salads.
So we put the turkey in the oven, and then headed to the garage for some music. It turns out that our average extended jam session length is about the same amount of time you're supposed to wait between basting the turkey. Skip was on the drums, I was on the vibes, and Jason was on the bass, achieving a nice balance. We wrote a few songs, ate a massive amount of food, and then passed out. Good times.
Thursday, November 15, 2007
Monday, November 12, 2007
binary solo! (flight of the conchords)
I can't possibly do this band justice with description, so I'll just post links.
The humans are dead
Business time
Angry dance
The humans are dead
Business time
Angry dance
Friday, November 9, 2007
How to be in a band
I've been in rock bands on and off for the past 10 years. They were all fun, and all dysfunctional, but in predictable ways. Looking back on these bands, I've noticed some patterns and some lessons I wish I had learned earlier:
1. Decide early whether this band is a collective or all about one person.
There are benefits to both types of band. If one person is running the show, the burden of writing songs, booking gigs, footing the bill, and doing marketing is mostly on that person. The rest of you just have to show up and play. Or if it's your show, you get complete creative control. If it's not your show, sometimes you do have to put up with hearing the lead person talk excessively about how awesome they are, and about their various relationship problems (which usually end up inspiring the songs). Sometimes you have to put up with getting your opinions carelessly vetoed. But you get to be in a band for relatively little cost in terms of time and money.
In a collective, everyone gets a say, and everyone is treated equally. Generally you all split the costs, and do the songwriting and marketing work together. This is more free and equal, but the downside is that everything takes longer. Writing songs as a group that actually congeal as distinguishable songs (and not 20 minute jam sessions) takes a long time. Also, you are constantly negotiating and compromising. And if your collective is comprised mostly of hippies (as collectives tend to be), the process of making any decisions (where should we play out? what kind of music should we make?) will take forever because no one wants to be the leader.
2. Create bands with people who are at your skill level.
You don't all have to be musical super geniuses, but being at the same level will allow to write songs you know the rest of the group can play, allow you to improvise more easily, and will allow you to avoid fights. If one member of the group is much more technically advanced than the rest, there will be an inevitable fight about this person being an uninvolved snob. If one member of the group is technically less advanced, there will be an inevitable fight about this person not improving. It's much easier to avoid this situation altogether.
3. Expect to argue about the band name for about 6 months.
Choosing a band name is hard. Finding a name that everyone doesn't hate is harder. But hold out hope, and avoid compromising on a mediocre name that nobody likes but nobody particularly dislikes.
4. Don't burn out
When preparing for a gig, it's easy to want to practice many times a week, for hours each time, but in doing this you hit diminishing returns too quickly. It's fun to sit and noodle for hours when you are writing songs, but when you're rehearsing a set list, it's better to limit practices to roughly twice the length of the set list. I personally don't find it useful to practice more than twice a week, and would find once a week ideal. I also don't like to have a practice the day before a show. Also, it's important to give yourself time to continue to play and create music on your own, so you can keep the creative juices flowing, and make sure you get some musical cross-pollination.
1. Decide early whether this band is a collective or all about one person.
There are benefits to both types of band. If one person is running the show, the burden of writing songs, booking gigs, footing the bill, and doing marketing is mostly on that person. The rest of you just have to show up and play. Or if it's your show, you get complete creative control. If it's not your show, sometimes you do have to put up with hearing the lead person talk excessively about how awesome they are, and about their various relationship problems (which usually end up inspiring the songs). Sometimes you have to put up with getting your opinions carelessly vetoed. But you get to be in a band for relatively little cost in terms of time and money.
In a collective, everyone gets a say, and everyone is treated equally. Generally you all split the costs, and do the songwriting and marketing work together. This is more free and equal, but the downside is that everything takes longer. Writing songs as a group that actually congeal as distinguishable songs (and not 20 minute jam sessions) takes a long time. Also, you are constantly negotiating and compromising. And if your collective is comprised mostly of hippies (as collectives tend to be), the process of making any decisions (where should we play out? what kind of music should we make?) will take forever because no one wants to be the leader.
2. Create bands with people who are at your skill level.
You don't all have to be musical super geniuses, but being at the same level will allow to write songs you know the rest of the group can play, allow you to improvise more easily, and will allow you to avoid fights. If one member of the group is much more technically advanced than the rest, there will be an inevitable fight about this person being an uninvolved snob. If one member of the group is technically less advanced, there will be an inevitable fight about this person not improving. It's much easier to avoid this situation altogether.
3. Expect to argue about the band name for about 6 months.
Choosing a band name is hard. Finding a name that everyone doesn't hate is harder. But hold out hope, and avoid compromising on a mediocre name that nobody likes but nobody particularly dislikes.
4. Don't burn out
When preparing for a gig, it's easy to want to practice many times a week, for hours each time, but in doing this you hit diminishing returns too quickly. It's fun to sit and noodle for hours when you are writing songs, but when you're rehearsing a set list, it's better to limit practices to roughly twice the length of the set list. I personally don't find it useful to practice more than twice a week, and would find once a week ideal. I also don't like to have a practice the day before a show. Also, it's important to give yourself time to continue to play and create music on your own, so you can keep the creative juices flowing, and make sure you get some musical cross-pollination.
Thursday, November 8, 2007
Practical Password Formulas
It's hard to remember passwords, so many people use the same passwords for everything: email, bank accounts, etc. The problem is, if a hacker gains access to one of your accounts, he now has access to all of your accounts. Some people save all their passwords in a file on their computer, or on a piece of paper on their desk to make the passwords easier to remember- this is also bad because there's only one location or file to compromise, and then a hacker has access to all your accounts.
So how to get around this? Have a password formula stored only in your head. A password formula is a simple algorithm that takes as input a key word (usually relating to the site you're trying to login to), performs some operations on this keyword to transform it, and the result is your password for the site. If you do this, then you only have to remember one piece of information: your formula. Mapping backwards from the password to the keyword should be so difficult that even if a hacker got access to one of your passwords, he shouldn't be able to deduce the formula and crack your other passwords.
Here's a (bad, insecure) example:
1. I'm logging into amazon.com, so my starting word is Amazon
2. reverse the letters and get: nozama
3. capitolize the second syllable: noZAMa
4. Turn the letter a into the number 4: noZ4M4
4. add the number of years you've used this site on the end: noZ4M46
5. ta da: your password for Amazon is noZ4M46.
This formula can apply to any site you're logging into. But please don't use it- it's very easy to guess. So how can you make your password formula easy to remember but difficult to create? This is a hard problem, but there are some helpful strategies.
Strategies to avoid:
- Don't allow dictionary words in your passwords. An alphanumeric password of length 8 has 36^8 = 2.8 trillion possibilities, but there are only about 300,000 common English words. That is a much smaller set of possibilities for a hacker to check.
- Don't use common transformations like turning the letter i into the number 1, or turning the letter e into the number 3. These are easily guessable.
- Don't simply translate the keyword into another language: there are password hacker dictionaries in other languages too (though, as one of my coworkers pointed not, not yet for Elvish)
- Don't simply add a number at the end of a word- this is also easily guessable
Some better ideas:
- Use word transformations based on your personal experience, free-association style (free-association is hard for other people to predict or guess). Example: one free-association type transformation rule could be "The site Amazon reminds me of the childhood memory of watching the Jungle Book", so transform the word Amazon into JungleBook. The rule is "x reminds me of the childhood memory y." Do you know what childhood memory Schwab.com reminds me of? I doubt it.
- Scramble the word so that no dictionary words remain. Example: switch all pairs of letters: JungleBook becomes uJgneloBko
- Funny rules are easier to remember. My old password formula had, as a transformation step, "Snoop Doggification"- I'd turn words like house into hizouse.
- Make sure your password formula has multiple steps.
And by all means do not enter your password into a website that claims it will tell you your password strength: what is to stop this website from archiving all the passwords people enter into a better hacker dictionary?
So how to get around this? Have a password formula stored only in your head. A password formula is a simple algorithm that takes as input a key word (usually relating to the site you're trying to login to), performs some operations on this keyword to transform it, and the result is your password for the site. If you do this, then you only have to remember one piece of information: your formula. Mapping backwards from the password to the keyword should be so difficult that even if a hacker got access to one of your passwords, he shouldn't be able to deduce the formula and crack your other passwords.
Here's a (bad, insecure) example:
1. I'm logging into amazon.com, so my starting word is Amazon
2. reverse the letters and get: nozama
3. capitolize the second syllable: noZAMa
4. Turn the letter a into the number 4: noZ4M4
4. add the number of years you've used this site on the end: noZ4M46
5. ta da: your password for Amazon is noZ4M46.
This formula can apply to any site you're logging into. But please don't use it- it's very easy to guess. So how can you make your password formula easy to remember but difficult to create? This is a hard problem, but there are some helpful strategies.
Strategies to avoid:
- Don't allow dictionary words in your passwords. An alphanumeric password of length 8 has 36^8 = 2.8 trillion possibilities, but there are only about 300,000 common English words. That is a much smaller set of possibilities for a hacker to check.
- Don't use common transformations like turning the letter i into the number 1, or turning the letter e into the number 3. These are easily guessable.
- Don't simply translate the keyword into another language: there are password hacker dictionaries in other languages too (though, as one of my coworkers pointed not, not yet for Elvish)
- Don't simply add a number at the end of a word- this is also easily guessable
Some better ideas:
- Use word transformations based on your personal experience, free-association style (free-association is hard for other people to predict or guess). Example: one free-association type transformation rule could be "The site Amazon reminds me of the childhood memory of watching the Jungle Book", so transform the word Amazon into JungleBook. The rule is "x reminds me of the childhood memory y." Do you know what childhood memory Schwab.com reminds me of? I doubt it.
- Scramble the word so that no dictionary words remain. Example: switch all pairs of letters: JungleBook becomes uJgneloBko
- Funny rules are easier to remember. My old password formula had, as a transformation step, "Snoop Doggification"- I'd turn words like house into hizouse.
- Make sure your password formula has multiple steps.
And by all means do not enter your password into a website that claims it will tell you your password strength: what is to stop this website from archiving all the passwords people enter into a better hacker dictionary?
Subscribe to:
Posts (Atom)