kvs

To this excellent list, I'd also like to add for each position posted: - The position's challenge of the week - The position's challenge of the month

This is one of the questions I ask every interviewer who interviews me.

I second that. Very useful program; pretty much any Microsoft product is available for download and use within the startup.

1999

Finding Flow: The Psychology of Engagement with Everyday Life (Masterminds Series) (Paperback) Mihaly Csikszentmihalyi

They can always sell it when the price is higher and lease it back.

Here is my issue with Things Sync:

Ever since they released their iPhone app customers have been asking them for sync. This being a productivity app sync is very important. Things for the longest time has had WiFi sync. It is a pain for two reasons:

1. I have to remember to sync my devices every time I make changes. This is not rest of my productivity apps work. My calendars sync automagically, my email inbox sync automagically, my filesystem syncs automagically, BUT my _productivity_ app does not!

2. Wireless syncing doesn't always work because in some environments they block the underlying protocol. Then I have to setup an AdHoc network etc. etc.

Failure on CC's part, IMO, is the lack of understanding, perhaps, of how important sync is to their customers. A good approach would have been to incrementally introduce support for sync, using Dropbox or MobileMe or what not. Instead what it looks like happening is they are building a whole new infrastructure that we will have to use to sync Things. My problem is that I already pay for two good sync solutions that work well. I don't see the point in paying for another one just for syncing.

Some how this all sounds like they are gearing up to develop a Things specific sync service. Last thing I want is another service I have to pay to sync my data. Already paying for Dropbox and MobileMe.

Why is this better than current approach? Is it because the current password remain unchanged until the user click on the link to reset part?

Recently I wrote "...user data is scattered among many web properties..." In essence it is "data scatter" Decentralization, in my mind, implies some sort of planning on where to put data (say to prevent failure or enhance efficiency...) but I don't think users plan.

Not necessarily all evil. Only text-book implementations require that one allocate individual nodes. Most implementations I have seen in Linux and FreeBSD can pool allocate the nodes. One can also pull off neat tricks to append link chains in these implementations to expand the list without the need to reallocate memory as needed by arrays.