The one who kept VLC free
welcometothejungle.co
welcometothejungle.co
This is an old interview that resurfaces lately (no idea why), and it was done by a HR company, and they did a lot of cuts. So, it is far from being technically accurate, and some parts are quite weird.
Clearly not for the HN crowd, tbh.
I dislike those kind of interviews, because they focus on one person (people-isation trend that we see also in startups) while VLC was done by a collective. It is sure that I'm the person that spent the most time on VLC and other VideoLAN projects than anyone else, but the work is a collective one.
Yes, it is true that I refused a ton of money to either sell the trademark and domain-name, or to include adware/spyware in the Windows installer.
(inb4 mpv, mpc-hc, potplayer is better...)
Thank you.
> ... or to include adware/spyware in the Windows installer.
Thank you a lot.
Sure, more money would be fun, but most of the people I know who have more money are annoying.
I think people will understand, if they read the article, that VLC is not a one-person project. You make it pretty clear in the interview.
VLC is one of the biggest open source projects _ever_. For it to maintain its quality over 18 years(!), it needs a leader with a vision and integrity. If it wasn't for an interview like this, people wouldn't know about the collective work that's happening behind the scenes. It's kind of a "necessary evil" to have a single person exposed to the media. And I'm pretty sure that VLC would have either died, been sold, or turned into adware if it wasn't for you.
So I think it's a benefit overall to have you as the sole focus.
Also thanks for the work on VLC, it's one of those things that's always been installed on my PC since the early 2000s. :)
Thank you and the other VideoLAN people
Pass my sentiments along to the collective. You all did great work.
I donate every few downloads. It is not much, but maybe enough people do it to matter.
Everything is mostly accurate, it is just the technical details that are not.
> Are there any interviews / documents where you elaborate on those topics?
No. Mail me :)
Thank you for your integrity. You walked away from enough money to retire on, out of a principled commitment to free quality software. That deserves a lot of respect.
Well mpv can play full-range videos with correct colors... ;) But for everything that isn't a full-range video, I usually prefer VLC. So thanks!
VLC can too, if compiled correctly with the libplacebo dependency. It shares most of the shaders with mpv.
I have a collection of subtitled videos from several years ago. When I obtained them, I watched them in VLC with no problems.
A year or two ago, I had the urge to watch them again, and I wasn't able to use VLC, because it could no longer display the subtitles correctly. This prompted me to start using mpv instead of VLC, for pretty much everything. But I'm still confused at how VLC could have lost the ability to play files it once handled perfectly.
I'm sure that if you still have those videos, the devs at videolan would be very happy to test them to repro that bug.
Also, if you know Vue.js, we need help on the new web interface of VLC. :D
Out of curiosity, could you give ball-park numbers?
[0] https://www.reddit.com/r/france/comments/736ghk/ama_je_suis_...
> (inb4 mpv, mpc-hc, potplayer is better...)
We can all enjoy libavcodec in our own special way, VLC is the masochist's lavc frontend, no kink-shaming. :- )
For example, he said he doesn't want to work on it full time because then he would burn out. His solution is to get a job not working on VLC and then just work on VLC full time.
But that just shifts the burn out to his official employer.
The main challenge is that most OSS companies don't have healthy revenue models for building B2C apps like VLC.
Customers refuse to pay which is why you see companies like Facebook that make the consumer the product.
If you have two products that do the same thing. A and B.
A respects your privacy and charges you $1.99 per month.
B sells your data and is free.
... the customers will all block to B and A will stagnate or just outright die.
I'm my own employer...
So if you downloaded a random video from the web and want to play it, you need to figure out what player you need to get in order to play it. Back then, a lot of stuff from the web was garbage, corrupted data, named incorrectly, .mpg files that were actually AVI format, etc. So once you found the right player, you had to worry about whether it would play properly with your PC's video card. There were flaky video cards and not all manufacturers had drivers that worked. Then, if the downloaded file was corrupted in any way, chances are the official player would throw its hands up and refuse to play it. Then, once you actually found the right player and verified that it could play the file back, you were still stuck with that player's horrible UI and feature set.
There were brave souls who would download these "codec packs" from super-shady web sites, to act as plug-ins to Windows Media Player, since Microsoft couldn't manage to ship working video playback out of the box for anything but AVI. God knows what else those codec pack installers put on your system.
And reliably playing video directly in your browser? LOL, that would be for another 10 years or so with YouTube.
VLC was a huge breath of fresh air! Maybe it's not so vital now, since proper video playback is "table stakes" on any platform. You can play any video now with one line of Python, right? But we did sorely need it decades ago.
For a long while now, VLMC has been a work-in-progress. It is a video editor being made by the same organization as VLC. I've been eagerly anticipating having this as a free/open source video editing tool that I can trust will be available in the long term.
I haven't seen any evidence that it isn't vaporware in a long time. So to those in the know- when can we expect to see VLMC finished?
No clue. Noone seems to care enough.
That's a real shame because there are lots of people who would really like to see a good open source video editor. Blender's built-in NLE works but it isn't as well-developed as VLMC could be (it's not the main focus of a 3D modeling program). I think Natron lost its main developer around that time as well. It's as if we almost had a broad revolution in open source video editing, but it petered out before hitting 1.0.
However, I don't believe I am skilled enough to be able to contribute to its development. Maybe someday.
> It was the fact that it came from the Linux world. So, when porting to Windows/macOS, the codec packages were part of the binary. It was supposed to temporary.
> This interview has cut that part.
I can speak about 3 of them:
1) the one mentioned in the interview, where, when VLC was ported to Windows and Mac OS X, instead of using the system codecs, they used the linux-way and bundled the dependencies. It was supposed to be temporary, and we would use the system codecs. This means that we control better our decoders and demuxers than other traditional players, and we did not require codec-pack. That became very popular for VLC.
2) The cone was a temporary icon (and a student joke), and it is sooooo weird for a player, that this became a huge brand because it is so recognizable.
3) The code was split in a lot of modules, to help speed up the compilation time, of the project. That made the code way easier to extend or port to other platforms. Indeed, for example, a lot of modules don't need a lot of maintenance, because they just work. And you can add features or new OS, without understanding the core.
There are 2 other technical reasons that helped VLC getting popular, but those were planned.
The interview is too short to explain all this.
Turn it sideways (5 degrees per major release) and keep the colors. Eventually it will be a Play Button.
I definitely remember the struggle with codec packs, then a breth of fresh air when ffmpeg codecs would solve most formats, and then just like that forgetting all codec issues by using MPC.
You should keep it.
My little girl has grown-up seeing the cone on our PCs and knows it's what plays the videos from my DSLR. She has asked to watch "cone videos", which means "videos of me you shot with your camera".
"That one papa. Movie" She is three plus, and can use a surprising number of apps. Communicates with me in emoji too. It is a hoot!
Any device she gets her hands on. If she sees that cone, score!
This is definitely the case, VLC isn't a very memorable name, but the icon is unique. I was searching the other day and after typing "traffic cone", one of the search suggestions was "traffic cone video player".
100,000 hours spent on programming the underlying features, 0 hours on actually making it usable.
That which is asserted without evidence may be dismissed wihlthout evidence - or, in terms which may be more paletable to some: lol, no u.
I use it every day for professional work, so, yeah no.
> 100,000 hours spent on programming the underlying features, 0 hours on actually making it usable.
Clearly you meant to say "0 hours spent learning it" there, right?
Bizarre keyboard shortcuts, awful ‘stay-on-top’ window behaviour when tabbing between applications. ‘Copy’ copies the inversion of the selection by default. Dragging a selection drags the rest of the image. The default behaviour of most tools is inverse to any previously established behaviour.
I’ve used Gimp thousands of times over 15-some years, in some cases for professional projects, and under the hood it is very good, but for many years it’s UI/UX made it literally the most bottom rung tool I would possibly choose to use for graphics manipulation - only when I absolutely had to. (I remember using Gimp in between the PowerPC -> Intel transition, where CS2’s performance on my Intel Mac Pro was worse than my iBook, and Gimp had a Universal Binary before Adobe did...)
I learned GIMP, through tooth and nail, and, knowing it well, I can still say through and through, especially historically, Gimp has been nothing short of a mess to use. I’m still not used to the idea that selecting and dragging something moves the rest of the image, not the selection itself.
I don’t want to learn to make an exception for Gimp as opposed to every other graphics application on the planet I’ve used since MS Paint, and I am certainly not the only individual criticizing particularly the glaring UX issues within Gimp.
Now, Blender, I’ve tried to take a dozen or so YouTube tutorials, and read dozens of articles on, and it just makes me want to hurl my MacBook Pro across the room every time. I was never able to accomplish anything serious in Blender the way I did with Gimp, which is why I put Gimp as the second worst UI/UX I’ve ever tolerated.
Unless there is an actual use bug, like things do not work, for high value software, it pays to just work the thing the way it works.
All improvements welcome, of course. And the interesting there is I hope they are good, not just a pretty pass, or rearrange. That takes time to remap. It is nice when it is time well spent.
I tend to give a wide berth to high value tools I get to use gratis.
Perhaps a better UX can be funded somehow.
‘Selecting selects the opposite’, is, to me, a use bug, among with many other designers and developers (a quick Google search confirms this).
I wish I had the time myself to go in and look at it, I really do. I’m sure it’s the same for a few folks. I love and support FOSS and I wish I could get behind GIMP like I can VLC.
But VLC managed to make itself a daily driver, even from this interview’s description, by being usable.
Gimp isn’t usable unless I’m going to dedicate myself to it. Functionally, there’s no reason I can’t use it and Photoshop interchangeably for the same task.
But when that same task takes me two or three times longer because I’m fighting with the UX, and design time is money, Adobe can unfortunately take my money for now.
Most of us have seen it all done a bazillion ways. I go all the way back to floppies, TV for a monitor and separate application and data disks.
Even a lame UI is presenting me with fast, potent tools!
If there is value there, the learning follows, then stuff gets done, next. It is hard the first few times. Easier after that. Almost no worry now. "How does this one do it?" K, next, done.
Hard to complain about pretty great tools I can use, modify, etc. for a song and small investment in how the tool works.
Or don't, right?
Adobe is happy for your money, and parting with it is worth more than some time sorting GIMP out.
No worries.
It is easy to create a tool that is specifically tailored to your workflow, because for parts that may tough to code, you manually take care of it. It becomes harder to generalize that tool for public use. It is harder still to add good UX that lets novices pick it up and begin working.
It seems easy, because the ubiquitous tools, libraries and frameworks have years of planning and experience to learn from.
In fact I've found that a lot of the professionals that refuse to go near the OS version are a hell of a lot less skilled than those that enjoy the benefits of the GPL'd versions.
Open Source apps are very popular in enterprise and professional circles.
Good, you will be happy to know that we're redoing the UI for 4.0, to fix UX issues. And the issue you mentioned is fixed in last version and better fixed in 4.0.
> but there seems to be no interest.
That is totally not true. We listen a lot to user feedback, but we are understaffed compared to the project size. It is getting better, and the biggest issues (like the one you mentioned) are all getting fixed.
I personally use IINA, because of its closeness to QuickTime UI, but I would imagine cutting / hiding so much usability would be difficult for VLC.
Can you tell me a little about your approach to this problem? I am super interested to hear what your thoughts on it.
Both :)
That said, VLC is still pretty good. It handles the various webm, mkv, mp3, mp4, plus whatever else I've got floating around in /home without losing a beat.
I’ve never opened media files from a command line, personally.
The ‘Swiss army knife’ approach is why I love VLC - the knowledge I can throw literally any file at it and it will just open is very satisfying. iTunes/QuickTime just don’t make that cut, even with Perian. (is that still a thing?)
Then how do you open media files? /s
This has been fixed in 3.0.7 (now, it increases a delay between successive failures).
VLC 4 has a new "player", but the behavior to fix this issue is exactly the same.
Tguillem implemented it in the player: https://code.videolan.org/videolan/vlc/blob/8fc06287a82f3f23...
And I implemented it in VLC 3: https://git.videolan.org/?p=vlc/vlc-3.0.git;a=commitdiff;h=d... https://git.videolan.org/?p=vlc/vlc-3.0.git;a=commitdiff;h=1...
This is implemented that way because retrying indefinitely may be the expected behavior (so we may not stop), we just avoid busy loops.
Btw, you should upgrade: http://www.jbkempf.com/blog/post/2019/VLC-3.0.7-and-security
And yes, of course there is interest (see the answer to your comment), but resources are finite and many of these projects have a very small fraction of the budget of even the smallest commercial operation.
When one gets little help and support and it feels like ungrateful demands are the only thing anyone ever contributes, it's not hard to see why one might just take the money and ghost.
A moment of respect for what jbk personally passed up to keep VLC free for us.
Comments like these never fail to bring a smile to my face. Someone who doesn't understand the project / issue decides it is easy and people should care about their personal grievances for a program they have not paid a single cent for
Just because it's an OSS project does not mean it's beyond reproach. I contribute free software myself and never once felt that I was above my users. I don't think the VLC guys are this way one bit, but your suggestion that I should not be talking about any of my minor gripes is ridiculous.
Just an analogy... Suppose on a nice summer sunday afternoon, you are walking or cycling on the country side. You come past a farm, and they have free apples on the side of the road that you can take with you. You pick one, but find that there is a bad spot on that apple. First thing you do is throw it against the front wall of the farm. The farmer comes outside and you tell him he should be more carefull in picking apples, it is easy to do that and won't take much time. Next week you are there again, and there are no free apples anymore.
I’d say there seems to be far less interest in improving, in particular, the UX (not even UI, but UX specifically) of usability disasters like Gimp or Blender.
I’m hardly ready to give Gimp another shot - I doubt, at this point, either of these releases will phase me or my use - I’ve simply been too disappointed with them both for more than 15 years of trying every new version, that they would probably literally need a new name to get me to use them.
I get a bad taste in my mouth when I think about Gimp or Blender - I unfortunately only have terrible memories of frustratingly trying to do basic tasks like make a selection or move an object around, to the point where I hear the word ‘Blender’ like I’d hear the word ‘Wasp’ - a little cringe of bad memories and something to be avoided.
I think it’s been almost 18 years of trying Gimp on a yearly basis on multiple platforms only to be baffled how rudimentary user-level functions are so fundamentally broken.
It must sound like I’m really bashing this software, and I’m not - I am appreciative of everyone who puts an effort into the important and often thankless work of FOSS - but part of the point of the FOSS movement is to demonstrate its feasibility to replace the equivalent paid software. This is also critical to its growth, evolution, and support. Firefox, VLC, and, to an extent, OpenOffice, have achieved this tremendously.
Due to the overwhelming complexity of the tasks it sets to achieve, Blender’s usability atrocities can be pushed aside. Not so with Gimp.
OpenOffice, for example, is somewhere in the middle, but both ways - the functionality and the UI/UX are both ‘good enough’, and I can hand it to my Grandma instead of Word and she will hardly notice.
I think the major point for me here, is that this is why there’s an age old saying that FOSS is made by geeks, for geeks, and we will never have ‘the year of Linux’.
That’s because usability these days is basically on par with functionality, because with so many frameworks and software options available, even if something half-works - people will still prefer the half working, usable thing, to something that works great and literally seems to change its mind on what function does what depending on some extremely random-seeming context.
Photoshop’s UI/UX is shit to start. I’ve often stated it’s a piece of software you must have someone basically show you the first time how to use - then it’s okay. That we have only this to really compare Gimp’s state of awfulness to, indicates, unfortunately, the depth of the shithole Gimp’s UX is really in.
I’m a UI/UX dev, and I find these pieces of software unusable, and, perhaps especially due to my education on the subject, actively frustrating to try to use - I pray to think what would happen if I gave, let’s say, my girlfriend, who does Photography and Graphic design, a copy of Gimp to replace Photoshop. She’d probably leave me for even putting her through the experience. /s
Again, though, OpenOffice vs. Word? No problem.
tl;dr: Usability has to go hand in hand with functionality in development, or there is a risk in the brand simply becoming a ‘bad name.’
If Gimp 3 might actually be usable, I’d personally go for a name change. I see people in IT visibly cringe when I mention it.
Very few UI/UX people do, yet they love to complain.
Unrelated, I found this statement (probably a mis-quote by the journalist) amusing:
> Because it was done as a student project, it is a very modular program and it is very easy to add a new feature.
So many people have found it challenging to make complex software modular and easy to add new features -- turns out, all we need to do is get students to create the projects, and then they'll be so! :)
Most of the time, they don't.
But lately, I created a company doing consulting and integration of libVLC inside mobile apps, and it is now employing most core developers of VLC.
Get a student who could be a really good software archidect create the foundation basically. I think it's a small amount of people who could do such things.
Certainly there are "students" with software engineering/architecture experience and/or natural amazing talent, but it's def not the first place I'd look!
I want to pay music artists better for their work than what I can with streaming apps, and I also want my entire library available offline on all devices.
VLC makes it not just possible, but easy and fun.
I got good news for you: next release has better audio quality (less resampling) and gapless playback. Plus more metadata support.
On desktop WinAmp is still the only one :(
There is a clear distinction between software written with profit in mind and free software. Behind the first one often there are people with agendas, strong opinions and marketing departments. They keep breaking things and force me to change my habits to their likings. The second group however, is user-oriented with no marketing department. I used to dislike vlc and its ui. After many years of experiencing software that breaks and pushes to change me I learned to love vlc. Even if it looks different. I don't care about this anymore. Actually I've developed a liking for it, let it break the UI conformaty. It works and is a solid piece of software and this is just brilliant.
One gripe though is that VLC gives a such a good impression that you get very disappointed when basic things like scanning your media library don't work. It catches you by surprise because the rest of VLC is so polished.
And for example the fact that there was no sleep timer until recently.
You'd think that an open source project so widely used would find plenty of help to implement such basic features.
"For a patch to get accepted, it needs to be maintainable as well as useful [...] Which means that we put a lot of emphasis on quality. So, with any code we get sent, we need to make sure that we can still maintain it in 6 months, 1 year, or 2 years… And that’s why we have a lot of features in VLC that some consider completely useless and why we don’t have others that some people think are important. It’s not because we’re stupid, it’s just that the code needs to be clean—if it’s not, you don’t get in."
(emphasis mine)
The ease and breadth of the ways to get content onto your mobile device is outstanding.
You can browse a Plex server, SMB share, cloud service, FTP server, and even point your computer's web browser at the IP address of your mobile device and drag and drop files onto the device, along with other options.
This has proven to be incredibly handy.
Kudos
If I could get everyone to take away five things from it they would be these.:
Keep everything as simple as possible, keep everything modular and as independent as possible, develop the features your users actually want and need, build on a solid base, and keep the maintainability of the project as a top priority.
Can you elaborate on the best way to ask for help in the VLC support forums? I've tried a couple of times over the years and literally got told to "stop asking for someone to do my work for me" and essentially go away..
I know the developers on the forum must get tired of the "CAN PLS HELP ME MAKE VLC. THANKS" type posts but mine were always specific questions with working sample code which illustrates smy issue and that I'd banged my head on for days without success.
I really want to improve the quality of our implementation but have no idea how to without some pointers from experienced devs.
Thank you.
It's hard and some devs are not very patient, tbh...
> I really want to improve the quality of our implementation but have no idea how to without some pointers from experienced devs.
Come on IRC. Really.
Got it - thank you.
Also applies to many other projects really, it has a big enough of a barrier of entry people's patience isn't nearly as eroded as in say public forums (or SO).
pretty great outcome :)
Like a boss.
We had JB on the Demuxed Podcast last year, and he told a lot of the early stories around VLC, it's much less edited than the version in this article. Here it is, with a full transcript: https://www.heavybit.com/library/podcasts/demuxed/ep-8-video...
Also if you get a divorce, you might not really be up to working on OSS for some time, which is probably part of changing wives. :P
This is presented as a positive. Here's the negative: poor maintenance. Virtually nobody knows, for example, how the official SMB plugin works, and there's no documentation.
Speaking about "best developers" (a section of the article). Since version 3, video playback suffered from one or more bugs on any of the 4+ machines I've tried it on. I'm talking about video playback, which is a core feature; I had to switch to MPlayer.
It seems to me that VLC is the software users love to love. I'm puzzled by the reasons of this; my guess is that products with a zillion of features have a certain charm.
Totally BS. We have 4 SMB modules, which ones are you talking about? (Windows, samba, libdsm (v1) and libsmb2). We spent a lot of time lately to fix that and improve it.
Sure, with the move to SMBv2/3 we had some issues, but that is normal because it required a new architecture, and they are almost all fixed now.
The documentation of those modules is at the same place as usual.
> Since version 3, video playback suffered from one or more bugs on any of the 4+ machines I've tried it on.
Since version 3, we activated hardware decoding by default, and that means it is very dependent on your linux distribution and your drivers. I think most of those issues are fixed now though.