Thursdays are back! [Update - July 15th, 2021]

I tend to agree @Vort but Lackys has a bit of history. He’s gotten less overtly negative in recent times but he doesn’t post unless it’s to make some kind of subtle dig on no beta or official release. :man_shrugging: Think he left a bad taste in most peoples mouths so he kinda tarnished his reputation a bit. To a degree that it’d likely take quite a bit to come back from I reckon.

But this isn’t about him. We got updates back! And it’s nice to get a steady stream of what’s going on behind the curtains.

11 Likes

Every time SN testnet tries to become decentralized, it either crashes, hangs or lose data. So no, it is not decentralized yet.

The question is “what prevents SN from working?”. No precise answer for it is a problem. Either technical or social.
If it was possible to evolve “v6.2” into working state with just bugfixing, then why not to do this?
If it is not possible without DBC, which means previous approach completely failed, then why no one talks about it?

7 Likes

The test nets haven’t lasted an impressive amount of time, yet. But nodes from home do connect and store and retrieve data for a brief time. So yes clearly there is an or are multiple issues but it doesn’t mean the concept is fundamentally broken anymore than it can’t last too long.

Seemed like the messages were still causing OOM issues if I remember correctly, which eventually bogged down nodes and then errors going everywhere.

I think your criticism is highly valuable @Vort, especially since you get your hands dirty with the test nets right away. Logging and improvements being done are being done to help solve those obvious but not so obvious problems.

9 Likes

And the concept of distributed storage in a redundant way is not new and works already. Usually used for large storage across few nodes but it is a standard filesystem that can be installed. I remember doing it in 2006. Have to look up the file system again. I want to say gfs.

Now SAFE is doing more than it did and in a more distributed way. As @Nigel says the basic system is working, just the normal run of alpha testing bugs being ironed out with some changes.

Also when MaidSafe do internal tests they are also on a distributed system in a more controlled way. To be fearful at this time when we are in alpha testing, with a basic working system that has bugs is premature at best.

If we had these troubles in systems/networks called beta testing or worse release candidate testing then I might be concerned, although not fearful depending on exactly the issues.

6 Likes

There have been a couple of people plus or minus at any given time who have been posting like how Lackys do. I’ve never really been troubled by it. Maybe others are, and that’s nothing for me to say anything about. But I personally don’t consider it an issue. If there are many people writing like that, then maybe there is something I need to hear there. If just the same few, not much to care about.

Anyway, I think @Vort’s view is sensible. What he expresses also tells me that the communication can be better.

Let me start with the question of why the testnets have gone down so soon, and why DBC now etc.

The AT2 payment system has had very little to do with the testnets hanging or data loss. DBC is not a way to fix that.
DBC was going in and now was a good time to take it in.
The AT2 system was implemented in a way that added to other existing causes of slow upload. The batching of payments could have been implemented for AT2 as well, but since we were going over to DBC we just skipped that.

Other causes of slow upload, or hangs or data losses have been about different things. Self encryption is implemented in a way that currently makes it slower than it needs to be. There were details in how adult liveness was handled that hindered query responses coming back to client. There were some issues with error handling/timeouts in qp2p. There have been more things, and there still are some there, and others are to come.

Batched payments and ops (remember, this is not a DBC thing per se), and different replication strategies (local first, to network when X) requires some refactoring of self encryption. Other work is ongoing with data types, how metadata and chunks are separated, what types are used and how. This also affects rewards, and payment system sets some boundaries in how that can be deviced as well. There is also lots of slimming down because of it, refactoring and some new code with data handling (also with messaging) which means this is a good time to review old conceptions and lay things down as they should, with all the information we have now. That includes adapting it to our imminent use of DBCs, instead of the deprecated system.

The work with finding out what causes the testnet issues is always ongoing, and it is in some places related to other work (such as that mentioned above) and sometimes less related. There’s also more or less clarity on what the cause is, where it is best attacked. It is not always a simple question of a single fix, since there can be many different parts contributing, and for each and every part there can be design decisions that influence in a gradual way.

Much more can be said, but I have to head out for breakfast, and I’m probably not around much to answer follow up questions now since I’m on vacation.
Anyway, hope that helps a bit @Vort.

37 Likes

If you do then you will see why. In good news streams he pops up with pure negativity to kill any celebrations.

It will, but negativity does slow a team and can kill it. Not we want all positivity, but as you do, we crave critique that is founded on something, anything. It’s just the uninformed negativity that seems targetted with the purpose of price manipulation or something that is useless.

We all have to cope with it, we are in the world, but we don’t need to stand still and accept it. I hope I do for a while but after more than a year of unfounded negativity then it’s time to say, enough!.

Hope that helps frame the dismissive response from me. The team are working hard and are dedicated, they don’t need that kind of critique, but as I say we all crave folk finding actual fault/questions etc. as that only improves us all.

20 Likes

Obviously 99.9% people dont understand what is going on MaidSafe · GitHub include me.

So they shout “what the fuxk is going on now?.” Well I understand but I think team do hard refactoring and do what SN needs (At2, DBC etc) since 2012. And I believe this attitude and funding to this attitude and knowledge of team.

So what can we do? Well we just need and have to wait until party start. That is all we need and what we can. Criticism? Haha well @Vort if you want to be a constructive critical advisor, go to github not this forum. That is meaningful.

@Lackys I think you better to sell it for you peaceful mind. Life is beautiful dont waste your time what you dont understand.

3 Likes

Seems like the difference between FUD and constructive criticism. The latter is appreciated, whereas the former is not.

Thanks for update! Sounds like things continue to move forward and T7 will be a belter!

8 Likes

I thought that only thing what was changed between v6.2 and your post

was removing of payment mechanisms.
So what were causing linear increase in processing time if not AT2 then?


GitHub contains almost zero community activity.
Some projects really use it for such purposes, but not this one.
Maybe it is just too early.

If you mean making PRs - it is good for small changes, not architectural ones.

2 Likes

You (or anyone) is welcome to point out issues there or ask important questions.

MaidSafe also has the developers forum https://forum.safedev.org/ where you (or anyone) can discuss any issues specific to development of the Safe Network

3 Likes

Is this an advice to use non-working feedback mechanisms?
I remember this report by the way.

By no means. Just a FYI of the options available

1 Like

Sadly that forum is almost moribund - Some years ago we had a debate to split this forum so the devs could talk amongst themselves - and that was fine - but this is now and I presume (always a dangerous move) that the real chat takes place on the internal Maidsafe Slack channels and there are very few titbits now as to how the techs are thinking other than Thursday updates - or when @dirvine @oetyng @StephenC @joshuef @bochaco etc post in here - honourable mention to @chriso though - lets hear more of you in here please :slight_smile:

4 Likes

The devs respond to anybody who asks questions though, or discusses anything.

Obviously when a testnet is out, this is the place.

Its a case of choosing your best option for what you want to do. I just made sure the options are known so there is choice.

Not sure why people jumped on the info as if I was trying to railroad discussions away. I was trying to help with options. I guess this is a contentious issue and any such suggestion is seen as a minimising attempt.

I will now crawl back into my hole and stay away from replying to such subjects

2 Likes

Indeed - was about to edit my post above to say - An observation rather than a criticism - but you beat me to it :slight_smile:

1 Like

The dev forum isn’t moribund, it alive and waiting for third party devs. We’ve got next to no third party development going on at the moment because the only API is the CLI, and the project has a very low profile. This will change, and I think it will be helpful for developers to have a space that is separate from what will become much more “lively” here.

9 Likes

Indeed and that was the original justification for establishing the dev forum

1 Like

It took what… 3 decenia to built the clear net?

Meaning you are delivering the Safe Network at lightning speed. Congrats, amazing, just…

giphy

WOW!!

Keep up the great work and just block those annoying users that try to poison your motivation with their negativity.

A few more steps to go!

8 Likes

Let’s logically step through this …

  1. If we didn’t believe it, we wouldn’t be here … As this is obvious, you know this, so clearly you are stirring negativity.
  2. If you didn’t really, in truth, like the project, you’d just walk away … Better things to do than attack something you have no interest in.
  3. So, you clearly DO have an interest in the project, yet you are trying to get people to fear that it’s never going to get released …
  4. The only reason that makes sense, is that you are trying to get people to dump MAID… So you can buy in at a cheaper price.
  5. Which means, your negative commentary … Is in truth very POSITIVE commentary for the project.

Lol… It won’t work. MAID holders have the strongest diamond hands of any project out there. You’ll just have to buy it at the current price buddy. :wink:

12 Likes

Yes I love to read those weekly updates. Last few weeks have been rather silent in my opinion. Understandable with the holiday season, still a strong on-going development on Github.

10 Likes