| Subject: Chat with Andrew Stein |
| From: Andrew Stein <steinlink@gmail.com> |
5:16 PM Andrew: figure out when yer leaving?
we need to have a going away party
5:17 PM me: 4th
hey
you think the core software could be ready by late August?
Andrew: software is never finished
5:18 PM me: you think something functional could be ready by then?
also, could you shoot an arrow at a target and have the arrow reach the target?
Andrew: if and when I release something, I'll let you know
me: if the arrow halves its distance each time, how will it ever reach that target?
5:19 PM I'm just kidding, I know the answer so shut up
anyway
Andrew: rounding error if the space is discrete
me: I need to have something at some point, you don't think you can make anything approaching an approximation as to whether and when you might have something similar to what we've talked about?
5:20 PM If you don't think you can, let me know, as I have to get our programmers to try to hammer out something rudimentary
Andrew: do what you will
Im researching a nontrivial search problem - some approaches work, some dont
5:21 PM me: right on
Andrew: i know the end point sortof
but you dont want what im building anyway
me: at any rate, I'd of course like to use your software for the project if you do end up putting out something stable and with just the basics
are you building something appreciably different than what we discussed a few months ago?
Andrew: im building exactly what i said i would build
5:22 PM me: okay, I do think that I want that
and that it would be very useful for our project, but if you don't think it works for what I need, I certainly understand
Andrew: ok, i hope we get to use it someday - i dont want to release something that is crap, and there is a reason that this hasnt been done 'right' yet despite the exposure
5:23 PM its a hard problem that needs to take into account alot of infrastrucutre that is completely speculative
how do oyu build a network on top of a use case tha tdoesnt exist yet
me: that's fine, just wanted an update
Andrew: ive tried many things - some work well, some are dead ends
5:26 PM me: now, are people forking your existing software at this point?
Andrew: i have 6 forks
me: right on
Andrew: but this is a research project for now
5:28 PM me: suppose I get our volunteer programmers to fork it from the point at which you most recently uploaded it, do you think they could viably turn it into a widget by which to send around data in the fashion I've described among bloggers, or would it make more sense for them to create something from scratch?
5:29 PM Andrew: i dont know what you've described among bloggers, but the answer is no
this is not "half a product"
i know what the product does
5:30 PM but there are many implementations that have various discrete sets of compromises
me: that's what I figured
5:31 PM Andrew: and again
i think there is a fundamental disconnect on thesis
you can support many of the features we've discussed in this sort of distribution network
btu the fundamental aspect of distribution is a loss of control, not a gain
5:32 PM you necessarily lose control over content in a distributed data store, period
regardless of the features oyu introduce, there is a fundamental abstraction leak in that the "content" you filter out is still there
you just arent repeating it
me: I know
5:33 PM Andrew: so no matter what, compared to a closed system
you are guaranteed to have less control over the content
but thats good
5:35 PM me: that's fine, let me know if anything changes or if you do end up developing something that could be of use to us
Andrew: when are we having a going away party for you?
5:36 PM me: whenever, I'm not on much of a schedule
Andrew: how about tomorrow or tonight, im off work
me: busy tonight
tomorrow is fine
5:39 PM Andrew: tomorrow it is