9

The Legacy of Layervault

over 7 years ago from , UI Designer

Dear Ladies and Gentlemen,

The legacy of Layervault has yet to become.

I've seen, I think about now, around 30 different posts of Layervault. All about what to do now, now its gone and what maybe can solve the problem it used to solve. It is even mentioned countless times whenever a new similar product launches or adds a feature which is comparable.

However I have yet to see a post that really looked back at the situation with a different angle and pointed out what is missing. Yes, Layervault has failed and yes, its gone down. But it did leave a message and it solved a vague problem that has yet to fully flourish its potential to solvation so to say.

What do I mean? I have thought of version control many times over the years even before Layervault came around. This as it's a real problem in not only the web design community, but everywhere where programs use big binary or even bigger ascii files. Of course this mainly involves the 3D community.

Studio's use up incredible quantities of disk space with countless versions of their products all in various stages of development. This has always seemed stupid and what has been the solution for YEARS? Naming convention.... yes thanks for this solution... and yes there has been svn etc, but none are of which are ideal.

Back to the point is was trying to make. Layervault did do a job which was too vague and at the same time too specific. It fixed a niche problem with a price which was too high in a world where the big companies are fighting to trow multiple gigabytes of storage towards your eyeballs at no cost!

Layervault only supported a few file extensions while the world is exploding with new design tools, or new iterations which use improved extensions.

Layervault restricted you somehow, something a version control system should never do. Thus it failed.

So what DOES a binary file version control system need to do? It needs to be extremely dependable, it needs to be accessible (Github is doing this sort of by supporting various image formats, which you can instantly review), it needs to be extendible and it needs to be extremely clear. All of this at a extremely low price (you work with big files, this is your job, you didn't choose the file size it came with, you should't be punished for it).

Four pillars basically:

Dependability means you need to have control over what happens and where your data is (your own servers!), without worrying about data corruption or loss.

Accessibility means that change between whatever files, needs to be easily made visible, plus these files needs be easily recovered/located if necessary.

Clear means a graphic user interface that fits like a glove. Git is made for the coding world. Git has a command line interface, this fits this community, yes like a glove. Whatever design community, it is a whole different beasty.

Extendible means that whatever extension and file format a business can come up with, they themselves can extend the platform so that their product is automatically supported. Open source extendibility makes this future proof.

All of this boils down to this point. Layervault did not fail. Its service has failed but it did put up a sign. Version control for the designer community is real. This problem should not be a part of a business, it should be a business on its own.

I think we can safely say the whomever does this correct, can put down a second git or Github. And That will be the legacy of Layervault.

thanks for reading,

12 comments

  • Tony GinesTony Gines, over 7 years ago

    Well put. You bring up some very valid points. I really think the biggest barrier to entry for me and my team was price. I really wanted to use LayerVault and just couldn't justify the price to accounting. We're already using dropbox business which offers some basic levels of version control and we couldn't see any more value beyond that. I think the real failure with LayerVault (and excuse my naivety here) was putting such a high price point on it. Sure, you guys were at stage one of the big picture you outlined, but it would've gotten tons more people into it and using it. I'm sure git and github didn't come out in it's current form in v1. You guys started the conversation and I wanted you guys to keep pushing it forward. I'm hoping someone picks up right where you left off. There is a huge need for this service and I'm willing to pain through v1, 2, and 3 if I can see a defined path forward.

    2 points
    • Daniel GoldenDaniel Golden, over 7 years ago

      Earlier on in it's life-span a LayerVault account cost about half the price it did near the end. I'm guessing they had to increase it to keep their head above water.

      0 points
      • , over 7 years ago

        I agree, however if the balance is tipped or wasn't correct in the first place, it will keep tipping over and over.

        I know when I saw the video of linus torvalds, which introduced me to the concept of git I saw the beginning of something very special and important. Layervault gave off a similar, but smaller and more vague vibe. The direction is clear and the real thing will overcome all obstacles eventually.

        It has to be said, that the goal was very clear what git should do. I do think this problem is more difficult to solve or point out.

        0 points
    • Dan MarrenDan Marren, over 7 years ago (edited over 7 years ago )

      Tony, I think that you may be confused. The poster of this thread did not work at LayerVault and has no affiliation with LayerVault.

      0 points
    • James Caruso, over 7 years ago

      "We're already using dropbox business which offers some basic levels of version control and we couldn't see any more value beyond that." That is exactly why when I was at an agency, we couldn't justify using it. Dropbox did versioning well enough. What's more, when we did try to use it, no one would go into LayerValut and use any of the features. We were all just so used to posting our designs and communicating on Bascamp.

      0 points
  • Du HoangDu Hoang, over 7 years ago (edited over 7 years ago )

    I think version control for designers ideally should happen within the files themselves, and not a secondary service like Layervault or Github.

    Imagine a PSD file where the history is preserved, and major versions are marked as such, and you can play back within Photoshop itself.

    Now that's a feature you can charge money for, not something that can adequately be solved by local storage and naming conventions.

    1 point
    • adrian ioadrian io, over 7 years ago

      Nice idea, but what would that do to file size and performance?

      0 points
      • Du HoangDu Hoang, over 7 years ago (edited over 7 years ago )

        I think file size and performance are solvable. Machine gets faster, storage gets cheaper. And I'm sure if some enterprising engineer decides to work of this problem, there will be clever ways for off loading the history in the cloud or something.

        0 points
    • Thibault MaekelberghThibault Maekelbergh, over 7 years ago

      Sketch kinda already applies this trough OS X's provided versioning system.

      0 points
  • Gabriele CirulliGabriele Cirulli, over 7 years ago

    Hey Dimitrie, as I mentioned I a previous post, I'm working on an open source version control system for design (which I'll reveal soon). I found your points really interesting, so would you be interested in having a more thorough discussion about this? If so, please email me at gabriele at impraise.com

    0 points
  • Jake FlemingJake Fleming, over 7 years ago

    I just use Apple's Time Machine or Dropbox if I need to go to a previous version of a file...

    0 points