Nice work!

Anyone will be able to create an AT (its creation is just going to be a "new type" of Nxt transaction).

It is likely we'll make it optional for forgers to "do" AT (at least for parallel chains) but for nodes that do they will need to "validate" an AT creation tx and then "execute" the machine code (and save the state as another "injected" tx).

Execution of each AT will be part of "verification" by all AT processing nodes (so nodes can't *cheat* with the *state* as those blocks will end up being rejected).

Personally I prefer "completed tasks" to hours (which is why I created CIYAM Open) - the problem with "hours" is that people can spend those just chatting or watching YouTube vidoes.

I envision that the way many of us ending working in the not too distant future is going to be much more like the CIYAM Open model (as it is far more efficient to "get results").

Fair enough then - with AT you can safely trade 1 Asset for another, with AT you can safely lock your savings away while earning interest and with AT you can buy a ticket every week in what will be arguably the world's most efficient lottery!

You can think of it has "being able to write contracts" - an AT will have basically all the characteristics of an "account" (so can send AMs and trade Assets, etc.) *but* this account is a "program" rather than a human being.

Thus - if you can read its code (or get someone you trust to check that for you) it can act as a 3rd party which "can't cheat" (I imagine someone will end up offering a service to *validate* ATs and *insure* them in the future to make it easier for non-tech people to use them).

They are perfect for things like "atomic cross chain transfers" and for things like trading Assets for other Assets without the need for complicated external infrastructure or "trusted 3rd parties".

Hope that makes it a "little clearer".

This issue is a non-issue as an AT with the ability to be sent Assets and to send Assets itself can easily act as a trustless trader between two parties.

Let's say that we have two account holders one who has ABC shares and one who has XYZ shares and let's say that those two users want to make a deal where for each 1 ABC sent and 2 XYZ sent the AT will then send the shares back to the other guy (if this doesn't happen by a "timeout" period then shares are returned).

Once established this AT can now "run forever" allowing those two account holders to trade ABC for XYZ shares at 1:2 (if later they decide to "change" the rate then this could be done via AM that needs to be sent by both users).

When you start to see what becomes possible through ATs you will begin to realise that there is simply no need to be doing other changes to the "core" of Nxt.

Yes - we are going to have to consider a few things about how AT and AE will interact to help with the above and with other things to do with the creation of a DAC.

Indeed the things that will be possible once we have AT in place are rather mind blowing.

Sure - display issues though are easily enough handled by clients.

I think we should consider raising the upper limit so that your problem will be easily solved via UI - technically there is no reason that the upper limit can't be that of a 64 bit integer (keeping in mind that Java doesn't do *unsigned*).

Yes - although with an appropriate AT set up for it the experience would actually be that you send Assets to the AT and then it sends you back the Asset being traded (so you send say Microsoft and get back Google).

Part of AT creation would include providing the appropriate "ratio" of such cross-trades as well as "limits".

In regards to the "limit" of Asset quantity I think we could consider making it much bigger than 1B.

Having "fractional" amounts is just not going to make sense in general but at the "client side" you could decide to show a "decimal" point somewhere if you wanted to.

Yes - my use cases already include AMs and it should be straight forward for an AT to deal with Assets.

Transfer of X qty of Asset A for Y qty of Asset B should be able to be done safely and atomically by an AT down the track so I don't think we should be considering any *core* changes for that.

Also parallel chains with atomic transfers via AT will be likely added.

Also - one forum member has stated he is working on translating my C++ prototype into Java (although he hasn't signed up to CIYAM or placed a "dib").

I think that the committee will need to work out *how* they want to approach the issue of funding tasks (whether it be by CIYAM Open or by a more traditional "contract").

JL has expressed concerns about this also (in terms of potential *legal* problems).

My "role" in the Nxt project is that of an "independent consultant" - the only "funds" I control are those that have been donated to the Nxt AT project on CIYAM Open. That does not function as an "employer" in any traditional sense but is a task based system where people "put in dibs" for a task that has a BTC "reward" (closer to say a "freelancer" model).

If anyone who has donated funds to the project is unhappy with this then they can just send me a signed message (using Bitcoin) to prove that they owned at least one of the addresses said funds came from and I will happily refund them.

Just to let Java programmers know that I am "not hiring" (it appears some miscommunication might have happened) so please don't PM asking me "for a job" as I am not in a position to offer that.

(sorry if that has caused any trouble for people)

My comments regarding "price" were not intended as "speculation" (but seemingly have been interpreted that way).

The market will decide "how much NXT is worth" and that value will be mostly determined because of the usefulness of the platform.

Perhaps people might consider "quoting this post" instead of the other but of course people are only going to quote what suits them.

Yes - it will take a while for this to "settle" (as any such change will).

Fees are always going to be a "political" issue - personally I think that the fee should be kept "high" as we don't want to *encourage* scams.

The main problem is "how do you trust" the chains reporting?

Basically "block chain tech" can only accurately report on "other block chains".