#Regarding S Class submissions

1 messages · Page 1 of 1 (latest)

ripe rain
#

I think the team is doing a manual review now. I'm not sure resubmitting all S classes will be ok.
I can suggest team provide an update on the next batch faster so the rejected task can be submitted again before the deadline.
Also Team is saying they have script automate read the sheet and update point on Nebb already.

#

What we can do now for tasks already submitted is they need to keep their RPC/explorer up until the end of SE
The team can do random checks and deduct points for that.
I don't think we can do this with IBC because maintaining old IBC is very difficult when chance is halted.

errant pecan
#

If task S takes too long to update, then at least the NEBB should be updated continuously. Currently, it is stuck at block 113,304

marsh prairie
#

Checking manually is the best way.

marsh prairie
agile zodiac
#

Yes resubmit works make more fud , and some ibc channel live at several days because it's small trusted periods, but i think check rpc easy can make a script

sinful blade
#

Prioritize low-submission high-value tasks beyond date order.

alpine berry
#

manual is the only way. the scheme you propose is a gift to those wanting to automate lots of accounts

#

that being said, why not involve community members in checking the simpler tasks? I wouldn't mind giving a hand if needed, and I'm sure others could be on board as well

supple vapor
#

RPC, IBC and Indexers should be evaluated and even health checked automatically

humble raft
supple vapor
lapis crow
#

IBC health check is not practical since IBC realyer costs so much toknes.

#

The rest can be done easily.

lapis crow
proper relic
#

Just to take some of the backlog out

errant pecan
supple vapor
errant pecan
#

example

[chains.packet_filter]
policy = 'allow'
list = [
  ['transfer', 'channel-3850'], 
]
lapis crow
tacit shell
#

Hire an intern for 2 months, probably the most efficeint way to tackle the problem.

  • Manual review
  • (s)he learns the project in the meantime
  • you get a boost on submissions check

I dont think automating is a good idea to implement. From a perfect world perspective, it could help, but it would cover maybe just RPC and partly IBC.
You can provide guidelines on how to format a submission however, for standard things :

  • RPC : provide the link, /status must be working, ...
  • IBC : channel must have new IDs, provide a tx id on mintscan that show a packet/tx got through that specific channel. Do we intend that the channel is always open and working ? maybe maintaining a working channel could give more points (coz now, it more a "fire and forget" task)
    And so on.

The difficult part, with the growing number of submission, is the deduplication.

#

Another approach could be "do a community check first" :

  • You dump all submissions in the spreadsheet, as soon as possible

  • You assign them a "community review" tag (not YES/NO)

  • People can comment on the quality/working task. Wait for 2-3 comments that point in the same direction (from recognized community members), then

  • "Team review" tag => YES/NO

  • Reward the people who take the extra mile to spend time validating properly, there are already some people in here who do that just to help

#

That way, duplicate (re)submissions are filtered out quickly, rubbish submission too. Spork can only validate that is shown to be working

elfin holly
lapis crow
#

I don't think it's so hard for ibc, rpc or indexer. All can be done by script. Especially for rpc and indexer, no need for human verification. As for ibc, it can also be done by script but more complicated.

#

Team just needs to review others, which is much less than these.

tacit shell
#

IBC : you just have to send a tx through the channel, if it arrives on the other side, ✅

alpine berry
high rune
#

We are aware of the team's hard work and we appreciate it.

wooden crest
tacit shell
#

that's why I added (from recognized community members)

verbal warren
#

🙏 Any progress on the S class review? Been waiting for 2 weeks now

lapis crow
alpine berry
lapis crow
lapis crow
alpine berry
lapis crow
delicate wolf
#

@here to be clear, we're looking for feedback. thinking that we:

  1. standardize the format for RPC, indexer, and IBC submissions
  2. participants review the sheet to see if their submissions is marked 'resubmit'
  3. "resubmitters" use the format to resubmit

this would only affect the above submission types, and from Feb 28 and onward (not Feb 27 and before). other kinds of submissions (like github and explorer) will not need resubmission.

pls review? and comment here or on the doc https://docs.google.com/document/d/1VDc0Xaa4mZPSoyG6tVbqFBZze6HFtnjNC2BZDIQi1TU/edit?usp=sharing

to be clear, this has not yet been implemented!

Google Docs
S Class Submission Standard
tired comet
delicate wolf
#

that's an important point
we'd only mark entries that are in a different format (because we won't be able to use automation to verify them)

lapis crow
lapis crow
# delicate wolf you will not!

RPCs: this endpoint is enabled by default once the node is started, just posting ip+port doesn't mean "support as service". It should have a domain and reversed proxy at least, and be synced most of the time. Otherwise, it is useless I agree on this.

#

at least a domain and reversed proxy. Also the node should be synced when checking. As for uptime, it can be monitored but a little late imo. Because the testnet is about to end in 2 weeks? or shorter?

verbal warren
alpine berry
high rune
#

Dear Namada team. I have a question. I definitely think that score for my task of providing rpc endpoints was given incompletely. It may be a bug, or there may be an error in automatic distribution. I would be glad if someone could review this issue. It would be appropriate if I could contact anyone on the point distribution system side. Thank u

civic dune
#

Sorry, of course, for my comment, but overall it seems like you create a problem for yourself first and then heroically solve this problem.))

When people say that the team is not keeping up with big data processing, it implies that someone is attempting to do so. As far as I understand, since March 8th, no one has even attempted to check the completion of community tasks.) And if you don't try - then undoubtedly, it's very difficult to check, and automation should be considered.)

alpine berry
#

was there talk of s-class submissions ending tuesday, or did I misunderstand?

spiral forge
#

feels like evrything is ended

#

so much of silence 😅

tired comet
eternal pivotBOT
#
katrin9253 has been warned

Reason: Posted a link

maiden marsh
#

652 🧐
you gave points for this dashboard, interesting point 🤔

alpine berry
delicate wolf
delicate wolf
tired comet
civic dune
delicate wolf
delicate wolf
civic dune
# delicate wolf can you expand on that? would be helpful to understand

Do we even need more comments for the current test network?)) Disregarding technical difficulties, since the test network doesn't imply stability.)) But there are many questions regarding the organization, starting from poorly thought-out tables, conducting IBC checks when the network wasn't operational, lack of initial automation, and ending with reward distribution, as well as the selection of genesis participants, whose Discords have just joined the group, while old, proven participants were thrown "overboard".)) Actually, it doesn't matter anymore; we're working with what we have.))) Just a tiny bit of unnecessary feedback for anyone.))

jolly void
spiral forge
#

@delicate wolf when we can expect a new batch from s class task ?

upper hearth
upper hearth
#

i provided my gist link in the evidence section. is that okay? or should i resubmit the submissions?

frigid coral
#

how does this plan sound?

  1. during today, i import existing submissions up to today's date
  2. i run the script that will check all those that already match the updated submission guidelines we posted the other day
  3. i add all (processed and unprocessed) to the sheet

that way everyone can see the state of their submission... no more wondering if it was received, or if it's in a future batch, or been checked yet etc. and also people will be able to see if they were already accepted so they know they don't have to resubmit

tired comet
marsh prairie
frigid coral
marsh prairie
#

@frigid coral It's really great. thanks for your effort. I think there are much fewer other tasks (except indexer, RPC, IBC) that can be checked in short time. isn't true?

frigid coral
#

yes i would say indexer/rpc/ibc make up about 80% of the submissions
rpc and indexer checks are easy to automate, ibc has been a pain point so far because there was no guidelines on how or what to submit so they can vary a lot

frigid coral
#

@here: i've updated the sheet to include all submissions up to today, after parsing according to the updated submission format guidelines (https://docs.google.com/document/d/1VDc0Xaa4mZPSoyG6tVbqFBZze6HFtnjNC2BZDIQi1TU). we are at over 1800 submissions so far

  • only rpc/indexer/ibc tasks were checked so far; explorers etc still need to be checked manually. so if your task doesn't fall into one of those three categories, you don't need to do anything further except wait
  • if your rpc/indexer/ibc submission was marked 'yes', the check was successful and you don't have to do anything further
  • if your rpc/indexer/ibc submission was marked 'no/bad format' please resubmit following the format guidelines above. (it doesn't mean you were rejected, just that the submission couldn't be parsed)
  • next focus will be on going through the unreviewed (yellow) submissions and comments on the sheet, and then the manual review tasks, to give people time to check the status of their ibc/rpc/indexer tasks and resubmit if needed
  • as always, there is opportunity to check the correctness of evals and make any updates

There were a lot of rpc submissions that were word-for-word identical with a http://ip:26657 address (see rows 950-980 for example) that, to me, look scripted, so i marked all that followed the same pattern as spam. but if this seems wrong, please let me know

marsh prairie
upper hearth
#

hey @frigid coral i think there's a mistake in my s class submission check. my rpc endpoint submission was submitted on 26 march (row 703) and it was checked as "working". but now, in row 1333, the same submission is checked as a bad format. please fix this

pk : tpknam1qzaqvxm6g4h7z34pgdqyzh3y4xevx072ednsfcujfza02knwrfkwqrgywah

tired comet
alpine berry
#

mann there are a lot of ip:port rpcs in that sheet...

alpine berry
novel bough
alpine berry
novel bough
#

emmm

alpine berry
#

we're speaking of s class submissions, not submissions of opinions

#

with this sheet I mean

novel bough
wooden crest
# frigid coral @here: i've updated the sheet to include all submissions up to today, after pars...

I am so skeptical of all those publicly opened RPCs with direct access to the host 😈 don't kill me but I would not count this as a working service coz of 2 things:

  1. you expose access to your server opening the node to all vulnerabilities
  2. it is enabled by default just when the node starts, so the only that is needed is to know your address and port (but cmon it is default 26657 for 80% of them)
    here is the recently updated list of opened RPCs, some of validators as well there: https://snapshots.liveraven.net/snapshots/namada/rpc-scanner.html
alpine berry
deft leaf
spiral forge
novel bough
#

Besides, I haven't set up SSL because the RPC data is originally public. Even if the signed data gets intercepted by internet service providers, what's the big deal? It was meant to be broadcasted anyway.

alpine berry
novel bough
#

All I can say is that for crew members, your advice is utterly nonsensical. Of course, a verification node should definitely set up sentinel nodes. However, this is just a test network, so I don't see any reason to fuss over this, it's just like everyone randomly proposing and voting.

novel bough
alpine berry
#

that's what I think anyways, you don't have to agree with me

warm cosmos
brazen violet
#

My post-genesis node uptime keeps showing 0%, why ?

verbal warren
upper hearth
#

the nebb leaderboard hasn't been updated yet?

alpine berry
brazen violet
bronze crag
spiral forge
jaunty bone
warm cosmos
warm cosmos
wintry veldt
#

how to submit s class task sir?

delicate wolf
delicate wolf
# civic dune Do we even need more comments for the current test network?)) Disregarding techn...

idk what "poorly thought-out tables" means, can you pls expand?
lack of formatting standards seems like a big miss, imo. that's what led to lack of automation. also i think that not making the contest Sybil-resistant was a really big miss. what do you mean about the reward distribution? also technical difficulties **should **be considered, imo--the software should have been more stable.

wintry veldt
#

hi @frigid coral can you please check my submissions? , Public RPC endpoint ht tps://rpc.namada.famnode.my.id/

tpknam1qz3eh8gfqgr2m44cfu3pg6sun4wgt75jwc8jvd85ax0wv24tfrrsc78gjet

thx