Scrame

[ still regurgitation. Response on the internet about visiting Tokyo. ]

Tokyo:
1) Akihabara is a place for hardcore geeks of all kinds, not just hardware. If you like collectibles, anime, consumer electronics, vintage music gear, stores packed floor to ceiling with every obscure console or home video game ever made, or just raw electrical parts, you can find it there. Along with cheap tools and strange gadgets, like keychain devices that interrupt IR remote controls for a cheap gag. The problem is how to find it. Check it out in the early afternoon most any day of the week. Most things start to close around 8.

2) The oft-discussed harajuku has a few areas, but definitely check out the freak bridge between the station and yoyogi park on a weekend, which is also a good time to see the park itself.

3) Shinjuku is OK, but its not too touristy except for how many lights there are. You might like Don Quixote’s, a bizarre outlet store for just about anything, but they have those in several locations; the same goes for Tokyu Hands, another unusual department store.

4) Shibuya is a must for a sightseeing trip. Check the harajuku exit in the evening, especially weekends, for an overwhelming flood of people, the area is full of stores, clubs, and restaurants.

5) Personally, I would avoid roppongi, but some people like that scene.

6) The city itself is huge, most of the stops mentioned here are just one station on the central line. If you had a better idea of what you wanted to see (museums, temples, etc), you would almost certainly find many more locations.

7) The train will bankrupt you.

Kyoto is the subject of several tour books itself, and with a good guide, you will never forget it. It is also a stark contrast to Tokyo. I have only been there as a visitor, so I am sure there are others who can give more detail. (See sanjyusangendo).

Miyajima (by hiroshima) is also cool, but dont let the deer get too close, they try to get in your pockets.

[An old draft I am publishing now.]

There are conflicting schools of advice about making a piece of software worthwhile.

Before I get into that, I want to mention that a lot of difficult problems _have_ decent out of the box solutions, where 10 or 15 years ago, you would be forced to invent the wheel. Most of these are abstractable service layers so that even the complex puzzle-solving bits can be swapped out (like with a rails app: you can swap between mongrel and passenger, or with Java/Hibernate: you can stub MySql with HSql)

Now, in terms of projects, we can find seemingly conflicting advice from Paul Bucheit. He wrote the prototype for gmail in a day, but then spent two years leading a team at google before releasing the ‘beta’.

I had some other points I was going to make, but in trying to find the endpoints, turns out there is a good middle: http://paulbuchheit.blogspot.com/2010/02/if-your-product-is-great-it-doesnt-need.html

It is important that you have a good idea and a vision for how it should be accomplished before you spend 2 years leading a team of people to launch it. If you have the good fortune to be in a position where you could just spend two years working on a single project.

Ideas are worth pursuing, but most of them aren’t worth two years. The simplest projects can run for years, so it is important to know that what you are working on is something that you can keep doing, and aren’t just running it for a quick cashout. Because that rarely works.

Its unrealistic to expect people to try lots of different ideas, but be committed to seeing ideas through.
If someone in your group can’t make a working example in a day or two, it probably won’t happen.
If your group is mostly people with titles and ideas, it will probably never happen, no matter how much money you throw at it.

In software, it is better to ask for forgiveness than permission. If you are bogged down in maintenance and operations, and can think of a feature or change or script that would make your life easier, do it.

It won’t always work, and sometimes will be rejected or won’t work. You can discuess it with opionated jackasses like me in meetings, or you could write the code and make it work.

Meetings and discussions are mostly unproductive and distracting.