Nxt Core Dev Q&A 11-04-2016

Bas Wisselink [19:59] @riker: @jean-luc you here?
Devin Luker [19:59] Saw  a reference to this in the NXT forums, heard some NXT devs were behind it, is this 2.0?  http://wavesplatform.com/whitepaper_v0.pdf
Testdruif [19:59] can't find a picture of "de pen is van je zus"
riker [19:59] hi
WhoCares FancyPants [19:59] Anyone up for a quick duel in Quake??? :simple_smile:
Bas Wisselink [19:59] @devlux: only one ex-dev involved, afaik
Devin Luker [20:00] So a fork? or clean implementation?
jean-luc [20:00] that dev has been ex for more than a year
Bas Wisselink [20:00] Anyway, let's kick off this third Core Dev Q&A
Devin Luker [20:00] K sry
Bas Wisselink [20:00] @devlux: no fork. Different system
[20:01]
Any questions for the devs? :simple_smile:
[20:01]
Then I'll butt out and play pacman
Devin Luker [20:02] That one was mine
WhoCares FancyPants [20:02] What do you think about making fNXT quantum resistant?
jean-luc [20:03] will think about that after everything else is done, but prefer not to change the crypto
[20:03]
enough other things are changing in 2.0
nxtswe [20:04] Have you started working on 2.0? If so, have you bumped into any problems / things you need to change since your proposal in the forums.
2615peter [20:05] How many ppl working on nxt core now?
jean-luc [20:05] yes, I have started, it still doesn't compile
[20:05]
for 2.0, me and ScripterRon, for 1.8, me, Riker, testdruif, Petko and Ron also
Martis [20:06] I have question: for childchains there are childchain owners, who pay fNXT to main chain forgers. Who is NXT sidechain owner, and who will pay fNXT to main chain forgers?
jean-luc [20:06] this has been answered before, I don't remember :)
[20:06]
there is no "sidechain owner" though
Martis [20:07] owner=creator
jean-luc [20:07] well, one thing I realized since starting work is that we don't need to have pruning and snapshotting implemented fully for the launch, just need to make sure what needs to be prunable is prunable by design
[20:07]
yeah, but the creator should not have any special control
2615peter [20:08] like MS?
Martis [20:08] so, who will collect all tx to the block and pass them to the main chain forgers?(edited)
jean-luc [20:09] anyone will have the right to do that, the default might be the forgers themselves, certainly for the NXT child chain
[20:10]
and for a child chain supported by some business, they can be doing that from some other account, not necessarily the one that created the child chain
[20:10]
but, the lifecycle of creating child chains is for a future release, first we start with hardcoded ones
Devin Luker [20:10] Not sure I like the collect and forward idea.  What happens if there is a rouge node that isn't forwarding?
Martis [20:11] well, I need an infographic to understand all this stuff completely :simple_smile:
Devin Luker [20:11] rogue
jean-luc [20:11] what does that mean? transactions are propagated to many peers at the same time
riker [20:11] I think we need to setup a bounty for infographic
Devin Luker [20:12] Sorry lagging...  Was a response to martis, so, who will collect all tx to the block and pass them to the main chain forgers?
Bas Wisselink [20:12] I have just started on infographics of all features :stuck_out_tongue:
[20:12]
give me the bounty, plz :smile:
Martis [20:12] yes, that would be very good thing, less confusion for everybody
Bas Wisselink [20:12] haha
[20:12]
it's on the TNSSE list :simple_smile:
[20:12]
I could make a thread for people to request specific ones
[20:12]
Might not be a bad idea
riker [20:13] @devlux: as long as more than 50% of the stake is honest we have no problem with rough peers
Bas Wisselink [20:13] images work much better than words
jean-luc [20:13] if you run a "bundling" process, which would be something similar to how forging or funding monitor runs now, you create and sign a ChildChainBlock transaction and broadcast it
Martis [20:13] not on features, but particularly 2.0 stuff. How it will work.
Bas Wisselink [20:13] agree, Martis
[20:13]
saves confusion
jean-luc [20:13] and if many such bundlers do it at once, the one that pays most wins
Devin Luker [20:14] Pays most in NXT terms or childchain terms?
josenxt [20:14] One transaction in a childchain will equal to one transaction in fnxt? I mean, the owner of the childchain will have to pay 1fnxt for every child-transaction? or will there be a way to group childtransactions?
jean-luc [20:14] if you already have such transaction in your pool, and receive another one offering the same fee, you just ignore it, if higher fee, replace it
[20:14]
if your turn to forge comes in the meantime, forge the block and get the fee
[20:14]
in any case, also share the transaction with other peers, similar to how it works now
[20:15]
in FNXT
[20:15]
group of course, in childchainblock attachments
josenxt [20:17] So the users of newer childhains will pay fewer fees thant the users of nxt childchain?(edited)
[20:17]
*pay
Martis [20:17] so, how it will work on childcains, where childchain inside tx fee is in childchain token?
jean-luc [20:17] yes, fee for each chain is in its token
[20:18]
we can try doing higher fees for other chains
Martis [20:18] so this native token fee goes to main chain forger?
durerus [20:19] why don't we need snapshotting and pruning right from the start of 2.0?
jean-luc [20:19] because it is complicated to implement
[20:19]
and when we just start, we will not have yet a bloat problem
[20:19]
we can prune retroactively those transactions later
[20:20]
will see when we get there
durerus [20:20] so we cannot market 2.0 as scalability solution?
jean-luc [20:20] of course we can, we just don't need to have everything in place at once
Devin Luker [20:21] Jean Luc, how stable is the addon API right now vs lead up to 2.0 do you have any major changes coming?
jean-luc [20:21] the Java API will change a lot, there is no way to avoid that
nxter [20:22] same that all ICOs out there, we can promote and market future features
nxt3d.net [20:22] Hello World :simple_smile:
jean-luc [20:22] there will be a lot of refactoring of the code
durerus [20:22] but without snapshotting it is not scalable
[20:22]
so what will come in 2.0 is one or two hardcoded childchains that will not be snapshotted?(edited)
Devin Luker [20:23] I get that 2.0 will be a major refactor.  I'm asking if there will be a major change between 1.8.x and 1.9.x  or if I can start building and not worry too much until 2.0
jean-luc [20:23] this will be faster to launch, and finish snapshotting in next release
[20:23]
between 1.8 and 1.9, no change really
durerus [20:24] but there will be snapshotting soon after? and snapshotting is really possible?
Devin Luker [20:24] Ok great, I have some addon orders to deliver this year.
1

riker [20:24] Clarification about API changes. The Http/JSON APIs won't change much in 2.0 what will change are the Java APIs used by add-ons
1

Devin Luker [20:25] Fair enough
jean-luc [20:25] yes, nothing has changed in the design about snapshoting, all I am saying is that we can launch before every piece of the design is complete
durerus [20:25] and what is the difference between pruning and snapshotting?
[20:25]
as i understood you prune by snapshotting
jean-luc [20:26] pruning is the removal of old transactions
[20:27]
snapshotting will be calculating some hash based on the current state of all objects that can't be pruned and are needed to validate future transactions
[20:27]
and recording this hash in the blockchain, and then new nodes downloading the chain need to have some mechanism to download the snapshot representing the latest state
[20:27]
and just verify it based on this hash, without re-processing old transactions which will be pruned
Martis [20:28] so it is something like the restore of everything if I'll download blockchain from scratch and want have running archival node?
jean-luc [20:28] they only need to reprocess transactions on the forging chain, which affect forgers effective balances
[20:28]
to verify that the forgers were legitimate
durerus [20:29] alright. thanks. and possibility to create childchains will also come soon after 2.0?
jean-luc [20:29] archival nodes will keep pruned transactions too, for specific child chains or for all
[20:29]
like they do now for prunable messages and cloud data
Devin Luker [20:29] Can you please for the love of god consider renaming forging to something that doesn't sound like a criminal act, please?  :smile:
3

jean-luc [20:29] this was discussed a few times before... nothing really stuck as a better term
Bas Wisselink [20:30] staking works
2

Testdruif [20:30] maybe a question for tenessee. there are 207 nodes with only 29 archival nodes. will there be some stimulus to increase these numbers?
Bas Wisselink [20:30] why not use what the industry uses?
Devin Luker [20:30] Yeah staking, minting, mining, crafting
Dave Pearce [20:30] also for me, staking works.
remix [20:30] damelon, just use it in outside communication
[20:30]
staking etc.
Devin Luker [20:30] Problem is when outsiders want to do due dilligence.
jean-luc [20:31] staking is like in peercoin, but I think their PoS is different
riker [20:31] What wrong about: Forging - is a manufacturing process involving the shaping of metal using localized compressive forces. The blows are delivered with a hammer (often a power hammer) or a die.
Bas Wisselink [20:31] yep
[20:31]
Forging normally used in the sense of "creating a forgery", Lior
Devin Luker [20:31] Because forging is also the process of creating something fake
Testdruif [20:31] staking has a negative undertone for me (being a vampire and such)
Bas Wisselink [20:31] Steaking
Dave Pearce [20:32] @testdruif...we've discussed it. Don't see the need for it yet, and every stimulus program idea has sort of fizzled out...like NSC(edited)
remix [20:32] language evolves. no consensus needed. use staking if you wish
Bas Wisselink [20:32] @testdruif about nodes: we need to do something there, but nodes require incentives, either from within the system, or without
durerus [20:32] will possibility to create customised childchain come before or after pruning and snapshotting?
Devin Luker [20:32] Look you need to understand, the folks who see this are risk averse.   Forging esp when talking to lawyers (which 2 are), just causes a gut reaction when they see it.
Bas Wisselink [20:33] So if someone can think of a great "outside the system" incentive, that would be great
jean-luc [20:33] will decide when we get there, depending on what is needed
Bas Wisselink [20:33] I've even considered striking a deal with raspberry pi and do Pi giveaways :stuck_out_tongue:
riker [20:33] We can suggest few terms and put for a vote
Bas Wisselink [20:33] But: costs and most people'd just use them for other things :simple_smile:
nxt3d.net [20:34] my question is related to the DGS and the after purchases operations, it is possible to add a parameter "returnURL" in the modal direct link, this returnURL will be used when the purchse is done to inform the source of the purchase about the sale and the transaction... thank you
durerus [20:35] @jean-luc: how much money did you expect to get for the development and marketing by ICO and how many fNXT did you plan to sell for that?(edited)
nxtswe [20:35] do we have enough funds to finish 2.0 ? I'm just talking about development here, not marketing etc.
riker [20:35] @nxt3d_net: please open a ticket on this here https://bitbucket.org/JeanLucPicard/nxt/issues
nxter [20:35] @remix we need consensus to change forging in NRS
remix [20:35] jean-luc, could you write a white paper for 2.0, or hire a technical writer or something. you are ignoring the speculator market. they want a white paper, perhaps even more than  working code
josenxt [20:35] https://nxtchat.slack.com/archives/general/p1460398724002355
Could the fees be selectively set for every childchain? Could some kind of "premium" childchains with fewer fees be created (and, potentially, IPOed? (for those ones who are up to IPOs)
 jean-luc
we can try doing higher fees for other chains
Today at 20:18

jean-luc [20:36] we didn't have any specific numbers in mind for ICO
durerus [20:37] riker talked about 500k dollars in marc's video. would this be enough for you?
jean-luc [20:37] the funding issue is not immediate
riker [20:37] I said that it would be nice to have 500k
[20:38]
Let's focus on technical questions folks
durerus [20:38] ok. at the beginning i thought you expected sums like lisk ICO
Dave Pearce [20:38] That would be kind of greedy......:grinning:
[20:39]
but useful.
jean-luc [20:39] yeah, the problem is that we have to compete with such well funded projects, even if not directly
Devin Luker [20:39] Did you guys change anything in 1.8x vs 1.7.x in the way peer connections are handled?  Since upgrading to 1.8 anytime i shut down my node it can take a helluva a long time to disconnect from peers if my network changed while the service was running.
Cryptkeeper [20:40] Nxt3d_net already mentioned the modal links, where can I find documentation for these?
durerus [20:40] you have any idea how we can manage funding? maybe explain better and let community vote on fNXT sell again?
jean-luc [20:40] hm, no, but are you using add-ons?
Devin Luker [20:40] Nope not yet :smile:
[20:40]
Only happens when I've changed networks in the past hour or two.
Testdruif [20:40] @jean-luc I also noticed this behavior with peershutdown
Dave Pearce [20:40] @durerus...we need to reach a consensus on lots of things to do with the transition to 2.0, funding being one of them
jean-luc [20:41] I have noticed some slowdown on start which might be related to having a security manager active when there are add-ons in use
Devin Luker [20:41] But I'm on a laptop so that tends to happen.
[20:41]
lag
Dave Pearce [20:41] Right now, all options are open, so we just need to keep on brainstorming until we do reach a consensus.
riker [20:42] Anyone having an unexplained slowdown, take a thread dump and send us for review. I'll post instructions on the forum.
jean-luc [20:42] could also be something with the newer jetty version and websockets... who knows
Dave Pearce [20:42] BUT: here is not the right moment to go into a full-on funding discussion. Forum....(edited)
Devin Luker [20:43] Is there any explicit GC?
jean-luc [20:43] no
Cryptkeeper [20:43] Where can I find docs about the modal direct links?
riker [20:44] @cryptkeeper: what ?
Devin Luker [20:44] I've noticed mine will run fine for hours then freeze solid for a good minute or two.  Got luck after I hooked it up to a profiler and found it doing some GC.
jean-luc [20:45] could also be doing the table trimming
Cryptkeeper [20:45] There is a new feature in 1.8.x: modal direct links AFAIR. How do it work?
Testdruif [20:45] @cryptkeeper if you are talking about calling modals or pages via url there is no up front documentation at this moment. But I will be happy to show you how you can find out all these things so you can document it :wink:
Cryptkeeper [20:45] Yeah, great. Thanks Testdruif!
riker [20:45] @testdruif: you actually wrote a nice PDF you can share
[20:46]
How do you like the desktop wallet ?
1

Cryptkeeper [20:46] I guess you dig up the link in the wallet by pressing F12?
Devin Luker [20:47] I like it, it's good work.  I'm worried abut JFX being alive long term though.
[20:47]
Did you consider electron by any chance?
riker [20:47] JavaFX is here to stay. Even improved in Java 9
jean-luc [20:48] I only hope it makes it to FreeBSD...
Devin Luker [20:48] Ok I might be out of loop, last I checked on it was last year this time and there was rumblings that oracle was dumping it.
Cryptkeeper [20:48] I have feedback from a helpdesk customer, who now can run the wallet the first time on his mac! So I would say: well done! :+1:
Devin Luker [20:48] Haven't looked since but we were going to do a kiosk based on it.
2615peter [20:49] you can keep javafx if you agree on installing the ask.me toolbar
jean-luc [20:49] the mac installer takes a lot of work
Devin Luker [20:49] Well that's retarded.
[20:49]
I mean ask.me
2615peter [20:49] lol
Devin Luker [20:49] Stupid lag
riker [20:49] I mean mac ...
Testdruif [20:49] uploaded a file

How to get the required data for url calls
634KBPDF Click to open original

Add Comment
1

Devin Luker [20:50] jean-luc have you looked at the packaging options in netbeans?
riker [20:50] @devlux: we are addicted to IntelliJ
jean-luc [20:50] don't use netbeans
Devin Luker [20:51] Might want to look at it for packaging purposes.  It's slick
jean-luc [20:51] well, prefer not to become dependent on an IDE for building the release
Devin Luker [20:51] but yeah I'm on the intellij bandwagon too, I know how you feel.
nxtswe [20:52] Is there any way to get the unconfirmed balance for an account where transactions not yet included in a block is included in the balance? (getbalance API unconfirmedBalanceNQT property does not include this)
jean-luc [20:53] no, unconfirmed balance is with outgoing transactions subtracted only, but incoming are not yet added
[20:53]
also other deductions like bid orders reduce unconfirmed balance
nxtswe [20:53] Ok, so fetching balance + unconfirmed transactions and calculate it on client is the way to go then.
jean-luc [20:54] if you ignore transactions other than simple payment
[20:54]
because in the next block the account balance may increase because e.g. an ask order got filled
nxtswe [20:55] ahh, i see, thanx.
[20:56]
Is there any guide to set up NRS dev environment for non-java devs? I did an attempt a year ago or so with Eclipse but never got it running.
Testdruif [20:56] intellij
[20:56]
5 minutes to set up
Devin Luker [20:56] jean-luc been meaning to ask this for awhile but keep forgetting.   If a national election were to be run over NXT what would the net effect be to the NXT network.  Would it be overwhelmed and collapse or keep chugging or what?
nxtswe [20:56] oh? cool! I'll try that.
jean-luc [20:56] I always compile from the command line even when using IntelliJ, so an IDE is not strictly necessary
nxtswe [20:57] I'd like to debug it sometimes, to learn the internal mechanics.
[20:57]
but I'll try intellij
riker [20:57] I once pushed a predefined IntelliJ project. Will try to refresh it.
jean-luc [20:57] how many transactions would that be... we support 255 per block
Testdruif [20:57] you can contac tme nxtswe I'll talk you through it
1

[20:57]
and then you can document it
[20:57]
:wink:
jean-luc [20:57] there was an idea to do such flood testing on testnet (or even production actually), but never got done
WhoCares FancyPants [20:58] 10 million transactions. But knowing people, everyone will vote on the last day at 23:00.
jean-luc [20:58] we currently have what, 2 million total?
Devin Luker [20:58] Probably a lot more than that.  Figure 15 machines per polling location 5 minutes per voter, ideally 200M people turn out, but ballots are a single message.
jean-luc [20:58] and this is for 2+ years
riker [20:59] @devlux: let's start with a village then a small city
nxtswe [20:59] We have 2,005,357 transactions it seems, congrats NXT ^^
jean-luc [20:59] yes, let's do the voting on that Greek island... what was it, 1000 people?
Devin Luker [20:59] @riker www.democracycounts.org is powered by NXT and just got funding.
[21:00]
Ideally a pilot project June 7th for the CA primaries.  So 10s of thousands.
WhoCares FancyPants [21:00]
>>>100000/255.
392.15686274509807

[21:00]
< 400 full blocks
[21:00]
Easy.
riker [21:01] We should better loadtest the network before that
Testdruif [21:01] it is just about generating transactions?
Devin Luker [21:01] Well validating them too
Testdruif [21:01] sending a message 1000000 times is enough?
nxter [21:02] 400/60 = 6.66 hours for the voting  :smiling_imp:
WhoCares FancyPants [21:02] Real vote would be better.
Devin Luker [21:02] I agree but the folks with money to throw at this want a pilot first.
Testdruif [21:02] yes, but can you vote multiple times with the same account? (in order to test)(edited)
Devin Luker [21:03] It's using NXT but not NXT's voting system.  Basically the machine has the account and generates messages to itself.
riker [21:03] Depends on the voting model
WhoCares FancyPants [21:03] I mean a real "testvote". With real "vote" transactions, not "message" tx's.
riker [21:04] What the point of using messages instead of the voting system ?
Devin Luker [21:04] i get that bcdev but there is a need to assure 1 voter 1 vote wihile still allowing it to remain annon.
riker [21:05] If you use messages how do you prevent double voting ?
Devin Luker [21:05] hueristics
riker [21:05] You trust a 3rd party ?
Devin Luker [21:06] no
nxter [21:06] @devlux give every voter a unique account and fund it with only one voting token and one nxt
Devin Luker [21:06] riker i have a 30 page doc that explains this it's too much to paste here
[21:07]
it explains why NXT voting wouldn't work for this, but this is me we're talking about and I think I'm rather concise on the forums, this one is verbose by my standards :simple_smile:
[21:09]
If you guys are interested and would like to read it I will gladly forward it along.  But consider that AFAIK there is no way to do a write in candidate for example.
riker [21:10] More questions or are we done for today ?(edited)
Devin Luker [21:10] Uhh would you like to let me know if that kind of traffic would be harmful?
riker [21:11] The network should be able to sustain this but I'll feel better if we do a dry run first.
Devin Luker [21:11] Was hoping you'd say that.
[21:12]
Messages are prunable too right?
riker [21:12] If you are really planing to pull such stunt why not endorse NXT as the engine for this ?
jean-luc [21:12] if those are just messages, there is almost no processing involved other than adding them to the transaction table, it is even less than a payment where you need to update recipient balance too
Devin Luker [21:12] Uh because when we originally designed it we weren't sure if NXT would actually do it.
jean-luc [21:13] how big are they? if less than 32 bytes, don't need to be prunable
riker [21:13] If you publish this now, we'll have 1000's of nodes by June 7th(edited)
Devin Luker [21:13] Depends on the ballot.
[21:14]
Fair enough and we will but I'm not the spokesperson.  I'm just handling the tech side.
[21:14]
Dan Wolf will be blowing in on Tuesday to chat about things.
[21:15]
We're pushing for this to become an international standard for secure eVoting.
25

Testdruif [21:15] can someone donate 1m nxt on testnet :wink:
riker [21:16] @devlux: I'm really not sure what to say ... can you bring up something solid to back this up ?
[21:17]
@testdruif: I will
Devin Luker [21:17] You'll need to define solid.  I'm asking you if you think the tech can handle it before they spend the money on doing it with NXT.
riker [21:18] @jean-luc: what do you think ?
Testdruif [21:18] @riker NXT-X3W3-4XM2-M2AW-DTY6U
jean-luc [21:18] not sure, it will be just the transaction volume
[21:19]
they are all sent from one account?
Devin Luker [21:19] Is there anything we can do about the volume?
[21:19]
Or I mean the impact.
[21:19]
It's not trying to disrupt the coin.
Martis [21:19] @testdruif: Are you going to make a script and send 1M messages on testnet? :simple_smile:
Testdruif [21:20] @martis probably yes
[21:20]
:wink:
[21:20]
altho I'm wondering if 1 node could actually handle that amount of transactions
Martis [21:20] Haha, I also prepared to code something to explode testnet :simple_smile:
jean-luc [21:20] the database index on the transaction table will have worse performance if a single sender account has so many records...
Testdruif [21:20] yep
[21:21]
I would at least go with 1 account per machine sending to a central account
jean-luc [21:21] we used to have this problem when every AE transaction used to have the genesis account as recipient
Devin Luker [21:21] Well figure about 5 mins per TX, polls are open for a day.
Martis [21:21] how many nodes forging on testnet?
WhoCares FancyPants [21:22] I can add 10 for a day if you need to.
riker [21:22] Very few
Dave Pearce [21:22] Better get forging on testNet, then.
WhoCares FancyPants [21:22] Or 30. :simple_smile:
Devin Luker [21:22] What is need to switch a live node to testnet?
jean-luc [21:22] those messages have sender and recipient, or just sender? if recipient is not needed, keep it null
Martis [21:22] can I run both nodes (main and testnet) on my raspi?
Devin Luker [21:22] Well it was supposed to be send to self.
riker [21:22] @devlux: a single property
nxtswe [21:22] @devlux: nxt.isTestnet=true in config
WhoCares FancyPants [21:23] @evildave: I don't want to run them for a week, just starting them 1 day before the test would be enough. VPS'es cost money. :simple_smile:
riker [21:23] @martis: perhaps
Devin Luker [21:23] But as long as there is a way to snag the messages then it should be fine.
[21:23]
They're signed, not encrypted and meant to be available publicly.
Dave Pearce [21:24] @bcdev...just ask people to spin up some extra testNet nodes, probably no need to spend lots on VPS
Martis [21:25] forging nodes
jean-luc [21:25] Nxt: forging your election results!
6

Devin Luker [21:25] Yeah another reason I'm pushing so hard on that name thing
Cryptkeeper [21:25] Can I run a testnet NRS simultaneously with my normal NRS?
Dave Pearce [21:26] And, yeah, you can run Nxt and testNxt together, not sure if RaspPi has enough power for the job, though.
[21:26]
Different ports
Bas Wisselink [21:26] I just snorted cola through my nose, @jean-luc thanks
riker [21:27] @devlux: If you bring us CA we'll change the name
Martis [21:27] well, I think we need to make another testnet blow , code some script and give it to tester to run
Devin Luker [21:28] @riker sounds fair.
[21:28]
Keep in mind this is a pilot run as an audit, but using the same exact tech that would be used in a live election.
[21:29]
Also once thing we're considering is putting nodes in each school and leaving them permanently.
[21:29]
}
[21:29]
In the USA you vote at schools
[21:30]
Anyways we'll open up a seperate thread Monday or Tuesday on the forums.  Sorry to hijack your dev chat.
nxt3d.net [21:31] can a wallet use a remote database located in an other machine or server? same as electrum or similar...
Devin Luker [21:31] Built correctly yes, that's all our voting machines are, same with mkoba.
jean-luc [21:32] just the database? haven't tried that...
nxt3d.net [21:32] yes, may lead to a light version of NXT
jean-luc [21:32] well, this is not probably what you mean
Devin Luker [21:33] Oh I think I misread.  You'd need the server sitting somewhere.  You don't want vairous and sundry connecting to your DB.
[21:33]
Right?
jean-luc [21:33] we are working on a light client mode for 1.9, which will be done using a proxy
1

[21:33]
making http calls to a peer with an open API
[21:33]
but it is not just the "database" that is on a remote peer then
nxt3d.net [21:34] the idea is a wallet whthout a database in it ... because it use an external one
jean-luc [21:34] who will provide that external db?
nxt3d.net [21:35] the trusted nodes i think
jean-luc [21:35] if you are talking about database accessed with SQL query over TCP/IP, no, this is not the way to do it
[21:35]
what you call database is really both the database and the java application around it, accessed using the standard http API
[21:35]
you can't expose your database to other nodes
nxt3d.net [21:37] there is a risk but the bitcoin electrum wallet work in the same way, the wallet don't need the bitcoin core to run because it use external servers where the blockchain and bitcoin core runnung
Testdruif [21:37] there
[21:37]
gonna send 5k messages to myself
jean-luc [21:38] ok, it will not be the database, but this is technical details
Testdruif [21:39] or maybe just 21 ish messages
nxt3d.net [21:39] i just share an idea :wink:
nxter [21:40] @nxt3d_net: there is already a light client for nxt out there: http://www.supernet.org/download.php
jean-luc [21:40] we will do it as a configuration option in the default client, this way it will always be up to date
[21:41]
and could be used while catching up with the blockchain
1

Martis [21:41] random node?
Testdruif [21:41] my node cuts out at 21 messages
jean-luc [21:42] will have to be random
Devin Luker [21:42] Speaking of which, did we just have a major fork?  My client is reporting I'm 75k blocks behind head all of the sudden.
Martis [21:42] or make option to choose manually, then I can use my own remote node
Testdruif [21:42] do you have a openapi testnet node around somewhere that I can abuse for a moment?
Dave Pearce [21:43] https://nxtportal.org/monitor/
nxtportal | Nxt Blockchain Explorer
nxtportal is a fast javascript block explorer for the Nxt blockchain.

jean-luc [21:43] I have block 13608157880320907690 at height 735157 as latest
riker [21:43] Me too
nxt3d.net [21:43] too here
Testdruif [21:44] same(edited)
Devin Luker [21:44] NM it's my fault, you guys changed where the blockchain is stored and I ran the desktop client.
nxt3d.net [21:44] 240 sec
Devin Luker [21:44] Linux funz
[21:46]
Is there an open source NXT blockexplorer ?
nxt3d.net [21:47] talking about linux i have some problems to connect to my node from the outside even with the firewall disabled and the IP  alone in the settings (removed 127.0.0.1 and localhost) :disappointed:
Devin Luker [21:48] Is it still an ec2 instance?
nxt3d.net [21:48] no a dedicated server in online.net
[21:49]
used for the nxt virtual world
Devin Luker [21:50] Oh sorry someone asked yesterday about ec2 and thought you were the same guy.
nxt3d.net [21:51] well well thank you all for this useful infos :simple_smile: i will back to my 3D @++
jean-luc [21:52] so, done with the q&a for now?
Bas Wisselink [21:53] That seems to be it, yes :simple_smile: