Letting out some steam: My Full Feedback on Current State of High Fidelity

#1

About to let some steam out and this is stuff that probably wouldn’t fit in a Community Assembly.
This is pretty much a rehash of

Most have been addressed, but there are quite a few things that still apply from way back then: so here is an update version for this year.

So prepare for yet another long epos. Some F-bombs may be present.


So what killed multi-con attendance numbers? Optimization. People were clearly having issues loading content (probably from LOD loss instead of networking as they sounded fine) among other things. and this drove folks away. While the event it self was fantastic, as soon as the ‘main carry attractions’ were over ( Talking to my self (very nice), and Mythbuster B Team chat), folks dissipated really fast. This was really noticeable as in the end we had much smaller number attending the Avatar competition show.

I recall during the “Talking to my self” skit, some big fans of James were difficulties still loading in content or seeing where the queue even started, but were coming in fine through the audio. It is sad to say but catering to just systems capable of running HMDs might be hifi’s downfall:


Now to my rant/ essay.

High Fidelity has very good tech on the network architecture but as an game engine it is barely barebones and as a Company, have moved away from an Open Source approach and more Closed.

But even after 3 years of pointing at it, the basics like Proper LOD optimization, and instead rely on stuff going missing, confusing end users more when their Middle-end PCs that work in VRChat or NeosVR get melted by High Fidelity or barely see shit. Of course you are talking about the Goliaths’ Unity vs Davids’ Garage Engine of High Fidelity… so that is a big ass leap, but come on, basics to improve usability for the average? Godot is fucking leaps above, and they are also an Open Source Engine. Even BLENDER Eevee is magnificent and they achived that in 3 years.

So what is so hard to gather resources to improve the engine? Other projects seem to do fine. Money? Resources? How do these other projects get so much then?

There are also quite a few things that grate my gears not related to the Optimizations among other things:

  • High Fidelity needs a community to succeed, and we have some seedling of it; but dismissing its main driving forces which were here in the beginning will drive them to prospect elsewhere. Give us the Soil and Water, and we will make something beautiful.

    • More focus smaller community events, fast build compos which ever. multicon was great, but was a failure since Hifi wanted more out of it. Maybe Focus on Nurturing the Existing community not trying to import other communities over. It is much cheaper than trying to pull off a huge event. Bingo is great, but creative endevours? That makes the platform UNIQUE from VRChat or Neos, or Sansar, on the fly building. Fast Builds, etc what ever. Quit relying on the energy of the existing community, as some of us tired trying to keep everything running. This, I am in agreement with @FlameSoulis
    • More Incetives to develop projects and attract prospective freelance content creators. Worklist was great, but it has also been neglected for years, (and probably just removed from use entirely.). There still isnt enough population to warrant some minor income from marketplace either. And with the removal of the Money tree, even less so from new users.
      - After Worklist was just forgotten about, most of this stuff was one-on-one, but even that changed lasted year; but even that excluded new creators that hadn’t established them selves. We need young creative blood. We have three - four handfuls of active hardcore old guard, but even that is dubious. We lost many great ones over the many years because of inaction and frustration; and some have even been banned from hifi domains due to them venting too much and cracking.
  • The sudden Clamping down on Community communication. Community Assembly is something, but its not; Its only a single time when our grievances are heard; but years of feedback has also been lost as they barely ever read the forums, things been more corporate. Road map is something of a good direction to communicate what is being worked on, but github also does that, and the canny roadmap is a bloody popularity contest of who can advertise something louder. We had the forum for ages, yet another suggestion place has been added. I have a feeling, even this rant is most likely lost in the wind.

    • Absolutely mind numbing number of alternations of communication methods, and non moderation of stuff. Telegram “exists” and was advertised, but it is a mess for communications. Discourse (this forum), Gitter, Worklist, and now we have Canny, and it just feels more is added for the sake of trying to communicate, but not succeeding that well. Please eliminate those that dont work and really focus on a platform. Github was working fine when issues were being responded; but it also feels like things are lost in the cracks constantly from all this moving around. You guys really need an ACTIVE COMMUNITY MANAGER that talks to us. Greeters are great, but any input feels as if lost through a hierarchy or not returned back upon at all unless you stick around asking the same question over and over again, wasting ones time.
    • Most things mentioned in the past have to be constantly reminded of in Assemblies or re-added to Canny. It is damn draining to keep pushing for stuff for years and not seeing anything in return. "Is it in canny" is also responded to is Yes: So it feels like even canny is only read by a few.
    • My observation is that (Current) HiFi employees have been go on other Social VR Platforms just to enjoy them selves because they really can’t talk about HiFi. It feels like its Fight club over there. “We dont talk about that”. Outside of Community Assemblies, Have we seen any fulltime-employee (but @Philip, occasionally between his meetings) outside of their working hours just hanging out?
  • There are absolutely Awful vulnerabilities and it took ages for them to release the info on the Filters. We requested these since I started doing Foxbombing. It took a year before filters, and another year for them to release the information for end users and domain holders; yet they keep info on some techniques used at multicon close to their hearts especially regarding Avatar Entities. The Transparency from High Fidelity is just missing, and for an “Open Source” project it is starting to annoy. Why must non-business stuff of the platform be held to secret. Or is everything considered business now? Its been happening for what feels like years now.

    • Public Bug Bounties. Give rewards for finding exploits and reporting them discretely. We have self proclaimed hackers here, lots of greys, but what about nurturing the white hats? And then actually doing something about them which helps everyone not just hifi hosted domains.
    • Running just any script: has been problematic for a while now, users have no control what script is being run on their systems, @humbletim ages ago made some client script whitelists, but never had any incetives to put them in the mainline branches. Canny: http://roadmap.highfidelity.com/feature-requests/p/client-side-script-whitelist/
    • Audio Injection is still an issue: there are no admin or user tools to see/ mute who is playing what sound, so folks spamming it and you cant even tell who said what. It is a wonder it has not been abused more akin to SL.. Only way to counter this is to boot the audio server, as clients can trigger audio injections via just console scripts and set locations to whatever, they dont even need collision sounds or entities. Most of the time the muting of the client also fixes or identifies them, but identifying them in a large ass crowd is impossible without the tools or randomly going through people to find them. Other solution is just reject any audio injection that is not voice based.
  • Marketplace: Ugh where do I even begin:

    • Images: why must is be such a fregging weird scale… Its fairly large too!
    • Scripts Opensourceness: Both Strenght and weakness of hifi. Scripters are mainly getting shafted as we cannot minify our code or obfuscate when we submit to the marketplace, they aren’t even obfuscated by the marketplace. We also cant use tools like typescript to if putting stuff on the marketplace. So @Judas, It is sorta like as if you have to upload all the original files (Blendfiles) model you made, everything including the Photoshop files of textures with all the layers intact so dont go out hifi-peen comparing. This is why marketplace is not even close to secure for scripters. Script kiddies currently can just “copy paste” code and claim it as their own.
    • @FluffyJenkins and I both have tons of marketplace items that will be “removed” within a few weeks, because we have to add a JSON file. Fluffy has even more items than I do. I am planning to just protest, and not update them.
      • This is trivial to do, but it takes time to add those. And for a platform that cannot keep users, its hard, very hard to put in the time to do menial stuff we aren’t really interested in doing.
      • Marketplace stuff could be just automated by Hifi; On creation of item and selection of key components but they haven’t even though of it. Why do we have update scripts we have made ages ago that worked well until now with a JSON file that for all points where to find everything for certification purposes. Its much easier for folks that dont care for certifications just to share links to scripts, similar to what @FlameSoulis and @Maki been doing, than deal with the marketplace, especially if our code has to be open source regardless of the certification.
      • Appification: Why must our scripts be apps and always on the tablet? Some functionality can be put behind menus, but the more scripts we put on the marketplace, suddenly the tablet becomes flooded with so damn many apps, all which are useful.
    • And speaking of Marketplace, the Update feature of just updating scripts of the marketplace item are still non-existant. We technically have to download the assets, update the script then reupload everything to update something. It is a damn hassle that could be streamlined with some clever web programming
    • And Speaking of Updating, Things break, my sign script has been broken after an update broke it years ago, so even if we put something on the marketplace and it worked for a nice while. No help was ever given and it stayed broken, and then removed from marketplace. It still relies on a bug with avatar attachments and scaling being fixed, but it never got addressed.

Without folks or money its hard to find the motivation, and without folks to spread money (as community is tiny), there is no awesome content, and without content being advertised, there nothing to for prospective folks to stick around for. It is a Chicken and Egg question, but NEITHER exist in big quantities now.

  • Engine work is lacking: LOD is probably the most important one. But it is something that needs to be done on the engine level. I have potential solutions but theyd just be stop gaps, as the main drain is Avatars. We need really need Custom LOD levels and perhaps an automatic one. This becomes even MORE important as Quest comes out. This should have been done already with the Android daydream stuff, but not many use that or able to use that, so mainline improvements should have been done… Remote connectivity and smaller hardware en-mass.
    • Engine still has long standing shader issue that have not been address 3 years after reporting them. Which would just absolutely confuse any prospective content creators. I understand that staff changes but jeesh. This is highly pertained to Material Optimizations and etc.
    • Rushing of Avatar Shaders before even Custom Avatar LODs are in place. Oh boy! the optimization nightmare for the poor engine team.
    • Material Entities are great in concept, poor in application: They are the last to load inthe scene and sometimes are not even applied prior to user coming close enough for the frustrum culling not to consider them.
    • GLTF: Years? when? Avatars? When?
    • Physics: Uf, Many issues here too big to list.
  • Avatar Soft Attachments are great, but they keep BREAKING. I can’t make use of a feature if it breaks constantly. Is it even being tested by QA? Is there a QA? It would make waves for makin customizable avatars.
    • Custom Blendshapes: When? 3 years after suggestion, nothing. ugh
    • Automatic atlassing. MMD Avatars are fun, but absolutely WRECK drawcalls. And apart of optimization should be to somehow take control over stuff that isnt normally controllable. It is a hard problem but it should be addressed.
  • Centralization: Domains Cannot be hosted on an independent network closed off the internet. There is quite a bit of stuff that relies on High Fidelity Inc and it’s services: So it needs quite a bit of work to disconnect the servers. How is this decentralized?
  • Network Stuff: While I praise High Fidelity for this, there is some odd choises regarding transmission bones. The Base Avatar bones should be prioritized over simulated bones, and there should be a capability for morph bones too that only are transmitted when the avatar loaded. No Scaling of Bones, etc, as once defined.
    • Scripts only know what the server tells them the client can see: Otherwise the scripts see nothing: This can cause interesting conditions with Entities.findEntities where not all entities in the domain are given, and instead entities that were loaded are only shown. This means inconsistant behavior.

The current unique things that High Fidelity has over the other platforms:

  • It can host crowds, but without people there are no crowds. It also doesnt matter much if most people are even able to render a crowd, how the heck are we gonna keep the crowds in before they just, dissapate.
  • On the Fly building: but the edit tools are bare bones. Everyone just sticks doing everything elsewhere… We will just have to wait until someone’s mad enough to do something about it.
  • Quick Scripting: This is High Fidelity’s current main double edged sword. and it Is huge, its much easier to get working and experiment with stuff. Everything is controlled via scripts, and we don’t have to compile libraries just do some basic stuff. The entry cost with some knowledge is much easier than jumping compiling Dlls for injected mods into Unity Games.
    • This can be expanded to allowing everything for avatar customization to procedural animation of avatar and even web cam tracking, if someone is determined enough.
  • Server Owning: Probably more specific to more technical folk, but its great to actually own our own domains.

High Fidelity should really get these nailed down better. Multi-domain stuff was also fascinating as a concept, but that was removed from scope ages ago.


But slowly, there has been stuff, which once made High Fidelity unique aren’t there anymore.

I came to High Fidelity because of the potential of what you can do with Avatars. I stayed due to opportunity to learn and maybe make some dosh on the side.

Removal of Face Tracking still absolutely stings me. That’s what made High Fidelity High Fidelity for me. Barely a shadow remains; someone technically could make a script available in the marketplace it most likely wouldn’t be free as it requires quite advanced knowledge, but honestly should be a core free feature. That is hard work however.

There still is tons of potential, and there have many things such as avatar customization (Detailed here in the past) that could technically allow for more security for ones own avatar (since only thing an fbx copy would take is base mesh, while External Blendshape API script access is non-existant without a custom client). But I’d be happy for Bone prioritization too I suggested above.

Avatar customization and custom avatar base-meshes, It would have been something unique to differentiate from Unity VR platforms and Sansar.


That, is my steam vented.

In short, I don’t have the patience to waiting for action on feedback we have given for years. This is probably the last long ass post, atleast for a while. As I’ve tried for years, and now I am out of tea.

I will persevere tho, but in the future I may not have the same vigor to push for stuff as I have prior: There are other opportunities, and I need to save my energy.

17 Likes
#2

Are you saying people stealing from the market are not criminals because its possible to do?

Removal of Face Tracking still absolutely stings me . this I agree with

1 Like
#3

No, I am saying that the point of entry to copying stuff from the market and just putting it elsewhere is low much lower with scripts. It is also harder to identify if something is copied or not when they can go around and obfuscate their code in their servers instead of in the marketplace, unlike “oh look this model looks just like @Judas work’”.

#4

Re protection of scripts, Is there anything that can be done to protect scripts that somone like you couldnt circumvent?

#5

Obfuscation can make scripts a harder to fully understand what is exactly happening youd really have to know how it is done, so somoene like me could do it, but thats from years of experience. A scriptkid just starting out tho? nah no chance they do it quickly. Given time, yes, but that takes effort.

It would be much harder to understand what exactly was happening where and would require a lot more time, and effort to get a part of the script you want to see what happens: by then they probably have the experience to write something similar on their own without copying stuff.

Right now though, with full source code its just: Oh look this is the function called foxbomb that does what I want it to do:. joink.

Instead of: Hmm. There is the api call todo change an avatar, but how the is x and y done? and how? wtf is function a? the fuck is b equals to, etc it is spaghetti to go through. If somoene really wants it, they will, but it makes the skill ceiling to do so higher.

Comparatively : It would be like packaging the blendfile and textures into an fbx.

Then importing it back to blender from Fbx file You can look at it, you can load it into blender and check the triangulated geometry: You can tweak it using blender, but you need to know how to do it. But you wont have the original.

You wouldnt don’t have access to the original photoshop or substance layers, or the original animation files to tweak them, but you have access to the finished textures, and youd need to know how to recreate or modify them in photoshop, or create anew.

With non-obfuscated scripts, its like having access to all the original files.

1 Like
#6

I’ve only been here for a month or so but I completely agree on each of these points. This frustration with the lack of developer communication and response/implementation absolutely alienates a community. HiFi has a small and devoted community for now but I do feel that it the development priorities don’t shift to match the community, there may not be much of a community to be had.

3 Likes
#7

Used to be whitehat, but at this point I might as well be a grey hat with some morality. Why? Because being the good guy sucked when everything reported was either already reported by someone else and is discovered to still be an issue, or issues are just left unsolved. The infamous mpassets should have been addressed upon initial discovery as it would minimize the potential damages involved to both creators and users, but instead Hifi insisted on waiting until their marketplace had even more items, started uploading avatars into it for personal hosting of users, and now has to sort that out along with some other nasty exploits I already reported (aka check your email, High Fidelity).

Called it when it was initially made, and yet is still an issue. I’ve had to post my Github issues as roadmap issues/feature requests and it feels absolutely stupid. Also, no offense to everyone, but shaders on avatars aren’t going to work if no one exists. It isn’t some magic bullet. Improvements to new user experiences, improvements for optimization (which I called out after the last stress test should have been the primary focus instead of doing server upgrades (which really won’t matter since we haven’t had 450+ users for over 6 months in ANY event)), and improvements to basic avatar features should be looked at first.

Also, I’m busy researching exploits and bugs, maintaining the Linux build guide, and trying to come up with new ideas. I don’t have time to soapbox my Canny stuff. Let’s look at a list of some good examples:
http://roadmap.highfidelity.com/feature-requests/p/make-avatar-certificates-readable
http://roadmap.highfidelity.com/feature-requests/p/port-certified-entity-protection-to-the-avatar-mixer
Oh look, some roadmap stuff about countering avatar theft before it became a rampant problem. Guess it wasn’t popular enough until now.

http://roadmap.highfidelity.com/feature-requests/p/allow-avatar-bone-priority-definitions
Automation should ONLY be used if the user doesn’t specify. If I know my tail and ears are low priority, I don’t need the engine to waste cycles figuring that out when I can just tell it. High Fidelity seems to love overdoing things and releasing it in a near-final edition when baby steps could be done and make each improvement be viewed as a sexy new feature or upgrade!

http://roadmap.highfidelity.com/bugs/p/ktx-textures-are-not-being-cached
This BURNS bandwidth for everyone! I’m not a financial wizard, but I can imagine the savings from this bug being resolved would be enough to buy everyone in the High Fidelity office a Happy Meal once a month.


To be brutally honest, most of what was said I’m in firm agreement with, and a massive chunk I haven’t been able to bring in my business partner from Second Life is strictly due to the incompetence of the platform, and not by my standards, but by theirs. This isn’t helped by me constantly asking myself “Oh, they wouldn’t let someone do that, would they?” as I then find about a dozen new issues, report to support, then find several months down the road it’s still an issue or the issue wasn’t looked deeply enough and is worse than I predicted.

On top of all of that, why should I waste my energy researching all this stuff anyway when it clearly is ignored? And now people are suddenly complaining about avatar thefts, grey-gooing of their domains, security vulnerabilities, users not coming in, creators not creating? Well, everyone, you’re welcome for reporting things that should have been fixed, being a good, law-abiding citizen of the metaverse, and told that “if it isn’t enough of an issue, it won’t get fixed.”

Not saying you were warned but… you were warned. Hope it gets fixed.

3 Likes
#8

On a lighter note, as per my discussions in world: High Fidelity has a bad reputation, and failing to improve the new user experience because “that’s how we want it” is a terrible excuse. If I had $1 for every new user coming in and leaving with negative feelings for the platform because everything was just a downhill spiral, I’d buy a yacht.

This isn’t helped by the community’s genius mentality of “Oh noes, bad reviews! Quick everyone, let’s share sunshine and rainbows and make all the bad reviews go away!” Oh, to make the platform rank go up more, to draw in more people on the disappointment train, only to then leave more bad reviews? Great job, people!

You want people to come in? Make coming into High Fidelity the most amazing thing ever. So far, all that has happened is we are literally putting new users in the closet and seeing if they ever come out. On top of that, let’s put them in front of one of the laggiest and buggiest things the world has to offer (cough cough mirror wheeze). Then let’s have them first meet Greeters who, even if they were in full body tracking, are thought of as NPCs (and yes, this has happened). Not to bash on Greet Squad (gotta keep an eye out for the old crew), but I do agree that something should be made more obvious that “yes, we are normal people who are staff.” I’m not saying use the old greeter tag (Though it can benefit from the lookatcamera property now shrugs), but there are ways of saying it. If a user first comes in, fresh new Interface.json and all, just have a script in the starting zone say “If you need help, a greeter at this desk can help you.” BOOM! Let the user connect that with the greeter on shift and go “OH, this person can help me. Cool!” 10 minute job for any coder depending on how you want to inform the user.

Any way you can make a user feel more welcome and at ease with the area should be a priority, and also realize that most people coming in are not some random person on the street. Actually, respect people’s intelligence and always include some ways to turn off the hints. Heck, there was a Help app. Capitalize on it! Improve it to actually give helpful information other than “Press W to walk. To look around, look around.”

4 Likes
#9

I see things differently

I see a few embittered users who spend their time scaring new users with tales of vulnerabilities foretelling doom

Damn one guy even explained to me how he now has a developer’s home address. That’s some scary crazy behaviour right there.

I see people trying to make names for themselves as griefers and trolls

I see people doing whatever they can to put off people right as they come in the door.

Why do they bother?

I would guess they are being paid but that’s the conspiracy theorist in me.

Maybe its just the snowflake generation sulking because they want everything done for them.

If HIFIs not for you leave. If it is roll up your sleeves and contribute.

I love developing my domain creating and selling.

Your happy to take the coin then stab the devs in the back

And I’m suppose to be Judas here.

3 Likes
#10

Vulnerabilities are important to fix. Because they, and the unhinged exist, when they shouldn’t.

Most of us have been discussing this on the side directly with the devs trying to get them fixed as they are potential vectors that need solutions. And many times they have popped up time and time again. And when such attacks occur to the new folks, they don’t need to hear stories from the grizzled vets about it when they WITNESS first hand it. I’ve seen many new user in TheSpot just never to come back again after a griefer attack takes place that moves them to the edges of the known domain, but from their perspective “everything just went black”.

I made the anchor script to combat that, but it is not a permanent solution.

What would you think, from a perspective of a new user come that comes in, and suddenly gets turned into a walking penis? will they stay? no, unless they like that shit: Oh look yet another troll to figure out how to break things here.

Or how about as a content creator, which I remember absolutely cursing about you domain getting deleted.

I recall you, @Judas complaining about it, and so I create the Build Tool Lock script. So of all people you know I wrap up my sleeves to try to fix and address stuff.

Or Rolling up the sleeves to get Blender avatar importing documented.
Or making blender plugin to make things more accessable for blender users.

But there comes a point where one simply runs out of sleeve to roll, and you can’t do everything your self. TAlking about stuff allows for other folks know about the plight and try to do something about it perhaps, or perhaps not and everything falls flat.

3 Likes
#11

I have always liked you Judas but in this case I find the need to slap you with the proverbial wet fish.
I have been sitting quietly lurking and reading, promising myself I wouldnt get involved, just because I am somewhat burned out there is no reason to spread my disappointment.
But seriously are you really going to dismiss Menithal as an embittered user because he has taken the time to outline some very serious issues?
He has spent years making things for people, he has done and incredible amount of work making tools for people and making this generally awkward platform somewhat more pleasant and useable.

Why do they bother???
Because they hold some desperate hope that their words might just be taken seriously enough to inspire someone of power to take an interest and fix some of these things that get mentioned with monotonous regularity.
They’re not griefers and trolls, most of them were dedicated users for years who have been burned out after countless instances of neglect and ignore.

We have all had our turn at pointing out some issues and bugs and complaining, you included.
So just because you are currently enjoying the Kool Aid is no reason to get on your high horse and belittle people who have genuine issues to mention.

What puts people off is what they experience in the platform and lack of response to issues, the forum doesnt put people off, in fact it is one of the areas that empower the users becausre they think they can use it to have their case heard, or would you begrudge them that?.

6 Likes
#12

Just i see coders describing to eachother how it works

oh hi @Adrian i rember when you use to come to hifi
and i have had numerous "community members " mess up my domain
and explain how easily they could still do it
it feels more like a threat of intent to me than helping

#13

describing to eachother how it works

Because thats how it works: You can’t create solutions to fix the problem before you know how the problem is done in the first place.

2 Likes
#14

well you obviously know more than the devs feel free to visit hifi again some day

#15

I probably do, given it feels like the devs only visit hifi once every month for a visit during the community assembly :stuck_out_tongue:
I atleast visit once a week.

4 Likes
#16

Happy Birthday Adrian. Miss you.

2 Likes
#17

So there is nothing hifi could do to protect scripts from you?
You are asking for somthing there is no solution to in this case.

#18

Menithal has been nothing but awesome. If it weren’t for his tools, neither my avatar nor my world would be in High Fidelity right now. Disregarding him as just a user looking to start something isn’t productive or accurate. Judas I think you’re trying to shut down what is a productive discussion. Everyone’s case here has been made professionally and politely in my opinion, and there’s no reason a social app shouldn’t have healthy discourse about its direction.

4 Likes
#19

Point being it’s the difference between locking one’s doors at night and leaving the door wide open. Sure, some people will find another way in, but most won’t put in the effort. Making it even just a tad harder can wipe out the majority of the people who would steal a script.

1 Like
#20

Im not disregarding him or his work. Im questioning the value in hyping up flaws in the system that only crazies would use . and the use of which i belive consitute hacking of a computer network which in the uk for sure is illegal.

if you leave your door open and i take your tv , i have still stolen it

2 Likes