Engineering Update: All Team Meeting, Condenser Split

in #steemit6 years ago (edited)

Much of our time last week was consumed attending an All Team meeting in Austin. We began that meeting with a retrospective during which we looked back over the last 3 months at what has gone well and what has gone badly with the objective of not making the same mistakes over and over again.

An example of one the things we felt we did right was dramatically increasing the amount of logging that all of our services do. This enabled us to roll out Hivemind in a much smoother manner than previous software releases. An example of one of the things we felt we did wrong was changing our priorities too frequently and spreading ourselves too thin.

Ads

After the retrospective we moved on to discussing our advertising program. We explored more ways to generate revenue through ads that didn’t compromise user security or adversely impact user experience. We also discussed how we might better optimize the ads we are already running, and how we could leverage our ad program to benefit all Steem app developers.

Steemit.com

We concluded our All Team meeting with an in depth conversation about how we should proceed with respect to the development of steemit.com (i.e. the Social Condenser application). Due to shifting priorities and unforeseeable events, development of steemit.com had slowed dramatically, but thanks to Hivemind and other stability fixes we now feel confident that we can, and should, make rapid progress on improving steemit.com. We explored the overall strategy we want to employ over the next few months, as well as which issues we feel should be tackled first.

Condenser/Wallet Split

We are still polishing the Condenser/Wallet split, but hope to be releasing that in the near future. Most of our testing is complete and we are putting the finishing touches on some UI/UX related improvements like making the top navigation bar disappear when scrolling down which increases available screen real-estate.

Soft Roll-Out

The strategy we are pursuing with the release of the new Condenser apps is a “soft roll-out” which will begin with the release of a stand-alone Wallet application to a domain other than steemit.com. At that time, a full version of Condenser will continue to operate on steemit.com. That means that anyone using steemit.com will see almost no changes. You can choose to use the new Wallet app by going to the domain, and at that time we would certainly appreciate everyone’s help in testing the new app with an eye toward ensuring that the user experience is as seamless as possible. The only change a user might notice on steemit.com is a warning if they attempt to log in to steemit.com with any key other than their posting key. The warning will look something like this (the phrasing is subject to change):

Screen Shot 2019-03-06 at 11.08.33 AM.png

Even this small change represents small progress toward making the proper usage of private keys a more pleasant and intuitive experience. That experience will only get better after the "hard" roll-out of the split.

Once testing of the Wallet app is complete and we are confident that it is stable and safe, we will release the stand-alone Social app. At that time, users will no longer be able to use their Active or Owner keys on steemit.com. The warning referenced above is intended to begin preparing users for that eventuality.

Cost Reductions

As we noted in our last update, most of the low-hanging fruit relating to infrastructure cost reductions have been completed. Until MIRA is complete, the focus will be on supporting other efforts like accelerating the release of the new Social and Wallet applications.

MIRA

Work on MIRA continues. We are working on a few bugs discovered through testing in our dev environment and hope to have a version for internal use in the near future which will dramatically reduce our costs associated with running steemd nodes.

Be sure to follow @steemitblog if you would like to see more engineering updates like these!

The Steemit Team

Sort:  

As this is a second question, what about SMTs? What about the previously posted roadmap regarding it? News about them were pretty thin lately.

We are definitely trying to get back to SMTs as fast as possible. It's not at all true that we are giving up on that project. The fact of the matter is that our work on MIRA is just more important now, and it's not possible to know exactly how long that work will take. It could be a week, it could be a month, we just don't know and aren't willing to pretend we do. Once we're done with MIRA, SMTs is high up on the to-do list.

MIRA is not more important that SMT's. As you can see, people are dying to build DAPS but your not giving them the tools. Those same people will leave to EOS or other blockchains. Mira is nice backend crap, but will not attract people to stay and build.

I'm not surprised people don't understand why MIRA is so critical, and that means we should do a better job with education, but MIRA is absolutely necessary prior to SMTs because it is a scaling solution. The cost of running steemd nodes is just too damn high. Period. MIRA will move the blockchain from RAM (expensive) to commodity hardware (cheap). This will dramatically lower the cost of running a node, which most dapp developers do. There's a reason why the only dapps people actually use are on the Steem blockchain, and it's because we solve the hard problems, not the sexy problems. People can waste their time building apps on other platforms, but the moment they get any traction, they'll realize they're screwed because their costs are going to go through the roof. We expect SMTs to dramatically increase the number of transactions happening on our blockchain (which are already extremely high) and so while MIRA will help every node operator (including witnesses) in the short run, it will also be extremely beneficial for ensuring a positive user experience once SMTs launch. It's solutions like MIRA that are actually doing the heavy lifting with respect to maintaining our significant lead over other projects in the space who aren't solving the REAL problems. They aren't capable of understanding what those problems are because they don't have real apps with real users operating in real time like we do.

Thank you @andrarchy ...

"... people don't understand why MIRA is so critical, and that means we should do a better job with education ..."

... for taking the time to educate us on the importance of MIRA. While my relatively small stake doesn't "speak" very loudly, I for one am all in favor of "doing it right the first time" rather than rushing through and dealing with all of the fallout of not ...

One of my favorite sayings (from an old "mature" engineer ...):

"If you don't have time to do it right, when will you have time to do it over?"

John Wooden, Legendary Basketball Coach of UCLA

Came here to ask the same thing. Seems Steemit are more focused on saving money than delivering crucial features. What they don't seem to realise is, if they deliver SMT's the price of Steem is going to skyrocket, therefore easing the financial burden they have right now. It's a catch twenty-two.

Appics is a great project and was depending on the SMT's being done this March. Now, they might leave to EOS because of these leaders have small vision and don't know how to set priorities.

Who cares if it's backend node crap is cheaper if all the projects are gone???

Steem will always be faster and lower cost than a general purpose blockchain like EOS. People are free to go to other blockchains, especially if their priority is short term financial gains that sacrifice long term usability. But those projects that do will fail because they will have to pass on their costs to the end user.

Who cares if some dapps lose patience and go to EOS? They'll be back when they see that EOS isn't fit for purpose yet. If they don't come back, I don't care because Steem needs people that have a bit of patience and can plan for the long term. Pretending everything is working, just to get a short term spike in the price is the wrong strategy, there's lots of other cryptos doing that. Perhaps more people should try using EOS, they've had a lot of problems and I'm sure that's not going to change fast. Most of their dapps are junk as well, there's only 1 that I like using, unlike Steem that has many good dapps now.

Well said. Which of their apps do you like using?

@andrarchy

Karma is pretty decent UI lately, mobile and supports Video. Getting much use in? AFRICA
Nigeria seems the future to me! 200 MIL smart, hard working people that have no job prospects and will have to devote themselves to becoming blockchain entrepreneurs.

People are free to go to other blockchains, especially if their priority is short term financial gains that sacrifice long term usability.

You should not be in PR, you should be behind a desk punching numbers and never have to talk to people. Your company released dates for SMTs launch sometime last year and shelved the project without an official statement or reason and this is what you have to angry disappointed people?

People trusted and have been waiting on you for over a year and just look at the tone with which you say the nonsense you say. Steemit should own up to it's shortcomings.

they were paying too much for servers to run the websites of all dapps. Cost reduction became prority for the company to surive. They are dramatically reducing cost and making some revenue from ads.

Once they are financially stable, they will work on SMTs.

Cost reduction is great, but you can work on multiple things at once. It seems to be taking them an unusually large amount of time to split out the wallet from the Condenser application let alone other things, which signals to me a couple of things:

  • The development team is lacking needed resources
  • There might be a disparity between skill level, during downsizing they might have let go more experienced (by proxy, more expensive) developers which meant they were left with only a few developers familiar with the more intricate underlying pieces. This means that there are probably pieces of the Steem architecture that realistically, only one or two developers can change
  • The Condenser app is messy, while it is broken up into individual components, it is still tightly coupled and difficult to work with. This seems to be a problem that many React applications share once they go beyond a few components

I think STINC is still in its infancy. From the outset anyone who can see things holistically, if Steemit had delivered on their promise of SMT's, they wouldn't have needed to experiment with advertisements. I can almost guarantee that they would make more than enough revenue from charging fees associated with the creation of a new token.

A great example of the demand is the work @aggroed and his team have done with @steem-eng which just launched its new DEX yesterday. Quite a few tokens have been created already, and while it is a sidechain, it delivers on the promise of creating your own token and it serves most use-cases. They're working on a consensus layer which will only put it further ahead.

The real problem here is: STINC is blind to their own community. There are so many capable and experienced developers who want to help Steem grow. Steemit Inc has more than enough liquid Steem to spend a few hundred thousand to fund community contribution and grow the blockchain. If Steemit let the community handle SMT's with financial incentives, they would probably be done by now.

Thanks for stating some insightful and valid thoughts. I note that @steemalliance is addressing your last point, perhaps differently than you'd recommend, but it is being undertaken.

Steem Engine is pretty cool.

Going forward, the number of dapps is going to increase at unknown, but probably accelerating rates, and poorly tuned nodes would have multiplied that burden obscenely.

It's impossible to state fees from SMTs would be able to provide more substantial funding, as there is practically no experience with such a product. Like Steem, SMTs are novel technology, and difficult to accurately price, or analyze for costs and benefits.

Most of us are convinced (I am) that SMTs are gonna be homerun financially, but if they come out flawed, they're gonna be a nightmare instead. I am encouraged that Stinc hasn't just rushed something out and started flailing away at bugs, flaws, and disappointment, as @drugwars has, for example.

I'm presently almost convinced that in the long run decreasing the expense of running nodes and encouraging dapps to run their own may actually be more critical to the success of Steem than SMTs.

Nodes are the key to decentralization, and if costs can continue to come down while computer capacity continues to increase, eventually extremely robust decentralization will be able to protect the blockchain from threats regardless of their source.

Nothing is more important than security from existential threat.

Lastly, I cannot judge your competence, but things are not always just what they seem to be, and it is often useful to consider the quality of optics with which one regards things. There may be extremely important reasons for specific undertakings being priority over others, and we just can't know without more information than is availed us.

Thanks!

Don't wait for Steemit, Inc to do this. Check out Steem Engine, a DEX marketplace created by the devs who made Steem Monsters.

We are committed to getting back to SMTs as soon as possible, that being said, I think Steem Engine is super cool and people should definitely check it out. Soft Consensus is Steem's secret killer app

SMT's are nowhere in their priorities right now. Steemit Inc is fighting for survival right now. We're not getting even a whiff of SMT's before 2020, and only then if crypto markets skyrocket.

I don't think they're still at the "survival" stage anymore, but that'd just be a guess after recent price changes and since they sell the same amount of coins monthly. About SMT's, @steem-engine lauched recently so people can get their coins going already and when SMT's come around to give those coins blockchain level curation, etc, we'll be waiting with a healthy distribution through markets and the way dapps are already distributing their tokens.

+1 Looking for info on SMTs and how to get an SMT name

I think the answer is above in that they must avoid spreading themselves too thinly which leads to not being able to complete cost related development that will help the underlying sustainability of the protocol and its infrastructure. They may have also been seeing the alternatives to SMT spreading and will likely have to take then into consideration when the time comes.

Posted using Partiko iOS

I think it also might be interesting in looking into giving authors which get a lot of people to see/click adds a share of the ad revenue. Would definitely get people to announce steem more loudly outside of the crypto space.

Without a doubt some folks would find them more palatable given that feature.

Perhaps too palatable.

I'd like to see how that actually worked out. It's a great idea.

Thanks!

Would also encourage authors to be more engaging in their own posts..

This would be an interesting change

Great idea! Thanks for sharing

Ads have pros and cons, as I'm sure you know better than do I.

Some negatives not commonly understood, however, are the potential for ads to be means to launder money, or to capture a platform and essentially hold it hostage. Advertisers wielding nominal stake with care can make a platform dependent on their funds.

While this seems unlikely on Steem, various malignancies can result even if the dependency is apparently less than total, or even seems minimal.

There are platforms today that only persist in operations due to their essential captivity to specific advertisers/communities. If certain emoluments feature in executive pay mechanisms, platforms that are losing money are still able to be abused, and such abuse concealed from investors.

There are very sophisticated players in the advertising market. Not a few of them are much different than what they seem.

I was happy to get out of that line of work. Do undertake specific care regarding covert attack vectors that advertisers may seek to employ to increase the influence they wield across Steem. It may be trivial, or only appear trivial, at the outset.

The game can be pretty savage, and the successful players are brutal. It is inadvisable for me to go into specific detail here regarding various hazards I became aware of prior to leaving that field, but some of them were existential.

Fair warning.

Yes, indeed. However, I think one should not just count the number of upvotes, but the number of comments (or maybe even better: comments on comments) under the articles to judge how many real readers an author attracts (bots aren't very interested in reading advertisements).

Appreciate the update.
1 - What is the reasoning for splitting the condenser into two separate applications? I don't think this is a good direction, and adds additional complexity by eventually forcing new users to extract their posting keys for signing in to the separate social app. This slows down the signup process and will cause a drop in user activation. This also will require all users to only be able to easily access the social app on a browser with a password manager, otherwise they will need to copy and paste this key. A SteemConnect login option may alleviate this.

While I agree that the wallet needs improvements, such as a more clearly designated send button and a cleaner address-book for contacts, splitting it out and forcing posting key only is a step in the wrong direction for UX.

2 - Could you elaborate on how you plan to use Steemit's ad program to benefit all app developers? Would you consider burning STEEM with a portion of earned revenue?

3 - At any time, has the Steemit UI development team considered improving the already built in promoted post system to feature these posts embedded in other feeds to drive visibility? This feature has been superseded by bid bots, which currently deliver cost-less post promotion, as rewards cover promotion payments. The Steem blockchain can capture more of this value by better optimizing the promoted post system.

4 - At any time, has the Steemit UI team considered adding a tip button to posts for sending a fast payment directly to a content creator?
To see an example of this, check out https://alpha.weyoume.io, currently operating on a STEEM based testnet.

All the best. -Haz

When are SMT's going to be completed?

You guys made a commitment to having this done in early 2019 and projects were literally depending on that deadline. Do you guys even understand the impact? Those projects, which would make STEEM look good, are now leaving to EOS or other blockchains.

Steem still has the best dapps and they can use this now https://steem-engine.com/

I was pretty upset about the going on of the past three months or so but, ultimately, Steem and Steem dApps are some of the best things to hit the internet since it's inception.

Despite some major errors and kerfuffles you guys seem to be working hard to right the ship. Good on you.

Question: can we expect to see communities by the end of 2019?

I appreciate you sticking with us through these challenging times. We're eager to get back to working on Communities and much of that work is already done. I want to see Communities in that time frame, and I think it can be done, but it's outside the scope of what we are currently working on so we aren't able to commit to anything. That being said, after completing the cost reduction measures that are our current priority (and which are going better than expected), Communities and SMTs are at the top of the to-do list.

Thanks for the really quick response.

It very encouraging to see this level of genuine communication and engagement with the community. I think that moving forward this could become as a positive point of distinction when comparing Steemit to other platforms within this space. Congratulations on the recent accomplishments and continued success.

Hopefully we can get some steem for watching ads, similar to what @partiko has done.

Posted using Partiko Android

Thanks for the update! It is great feeling that the platform one is deeply invested in is progressing inexorably into a spectacular future.

Thanks!

I am excited for you to be able to finish up work on the back end stuff!

Including ads was a good move. @partiko have taken that to the next level by rewarding partiko points for viewing ads. It will be interesting to find out the number of views these ads are getting in a few weeks time.

Loading...