[theora] NHW Project - some update

Raphael Canut nhwcodec at gmail.com
Thu Jul 19 05:10:18 UTC 2018

Hello Spork!

As your answer is highly interesting, I wanted to share it with the Theora
community (hope you don't mind).

You're right I should format the code properly, write more separated
functions, make it a library,... short, turn it into something

It's a lot of work, and I wanted to finish the codec (very high
compression) first... But any help is very welcome!

> I still see your codec being successful, especially in the electrical
engineering field, where system resources are limited.

Really thank you so much for your encouragements!!! But how can I convince
a company of it? I even suspect that companies don't even know that the NHW
Project exists, seeing the so few visits I have on my sites... There's also
that "the problem" is global now, a codec to become successful must also be
supported by web browsers, and I feel that clearly the giants of the
Internet are not interested in the NHW Project... How can I change that?
Would it take 2-3 years or more? Any opinion is really very welcome.

Many thanks again for your answer and for your precious time Spork!!!

2018-07-19 2:57 GMT+02:00 <sporkschivago at gmail.com>:

> Ralph,
> I would love to take over control of NHW from September to October, but I
> simply don’t have the time.   With our business we’re starting up, I
> hadn’t realized how much there was to running a business properly.   Money
> has now become an issue, and I need just a little bit more before we don’t
> have to purchase anything for a long time.
> But I also need to run all the low voltage wires, configure the Cisco core
> router, the Cisco wireless APs, the HPE 48-port gigabit managed switch,
> setup the VLANs, secure the server, setup the VPSes, learn KMS, etc.   We
> are so behind that we no longer put due-dates on the stuff that needs to be
> done.   I’m taking a small break so I don’t burn myself out.   I work way
> too many hours a day, but feel that I need to right now, in order to get
> this business up and running.
> If I where in control though, the first thing I would do is standardized
> the code, so people in the *nix world could read it, and it was nice and
> pretty.   I’d replace each tab with four spaces, use a consistent coding
> style, and I’d break the code into many more functions.
> Secondly, I would be using autotools, and work on making sure it compiles
> on Windows and Linux.   I’d try replacing declarations like int with
> declarations like u_int8, for example, because the size of an integer isn’t
> necessarily the same on all systems.
> Finally, instead of making it an executable, I would make it a library,
> with functions that people could easily call from their code, to encourage
> people to use the codec in their projects.
> Somewheres along the lines, I’d probably research more about codecs and
> see how to “register” one with the various OSes, so programs like Windows
> Media Player, VLC, XMBC, mplayer, etc could play a compressed video file by
> someone just double clicking.
> I still see your codec being successful, especially in the electrical
> engineering field, where system resources are limited.    Currently
> though, could I easily adapt your code into something that works with a
> Microchip MCU?   No.   What about an ARM processor?   No.   What about
> writing code for Android that uses your codec?   No.    I think for these
> reasons, you might not find a lot of companies, such as Intel, as being
> interested in your codec.    Results are nice, don’t get me wrong.   You
> have, to me, an extremely impressive “codec”.    But I don’t think a
> company would call it a codec right now.   I think they’d call it a
> program.    And I doubt many would look at it because 1) It’s not easily
> readable.    2) Functions are WAAAY to long.   3) There’s no easy way to
> call the functions from someone elses code (ie, it needs to be a library,
> and then maybe an exe that’s a viewer or can convert images to your nhw
> format) 4) There’s no documentation nor comments that explain what the
> functions do.
> So, what would I have to do to use your codec in my project?   First
> thing, fix the formatting.   But then, I would have to read through
> pretty much the entire code, comment how it works, and maybe break
> functions up to smaller ones.
> There are programs like DOxygen, where if you comment a function in such a
> way, DOxygen can automatically create documentation, in different formats,
> none the less (html, PDF, etc I believe), that explain what the functions
> do.
> You’d still need some example code, that shows people how to call the
> library to do different things (stuff like display_video.c, or
> convert_jpg_to_nhw.c).
> Finally, if I where you, I would read up a bit more on github.    When
> you implement a feature, you should push the commit, with a message
> explaining what you did.   When you fix a bug, you should commit.   You
> should NOT make a whole bunches of changes and then submit one giant push.
> If you submit the commits proper like, it makes it much easier to track
> down bugs, or to revert to previous code if something goes wrong, and for
> other people to help with your code.
> That’s a lot of work though, and unfortunetly, with my 20 month old
> daughter Chloe and being a new dad, owning a house for the first time,
> setting up a fairly large business (but only having three people to work on
> it, me, my wife, and my friend), I just don’t have any time available to
> help at all.
> I did start working on converting the code to make it standardized and
> properly formatted.    If anyone wants copies of what I did to continue
> my work, I would be more than happy to submit my changes to my clone of
> Ralph’s repository.
> Thanks.
> Ken
> *From:* theora <theora-bounces at xiph.org> *On Behalf Of *Raphael Canut
> *Sent:* Wednesday, July 18, 2018 1:58 PM
> *To:* theora <theora at xiph.org>
> *Subject:* [theora] NHW Project - some update
> Hello,
> Again I thank very much the Xiph forum members for their help and
> consideration for the NHW Project, really.I also realize I was a little
> unfair with Monty as just saying that Xiph.org can totally give some web
> space to document and advertise the NHW Project is already a very big help,
> it gives credit to the project,...!
> But the reality is that the NHW Project is a nearly dead project... Since
> 2-3 years so so few people took a look at the NHW Project, and I have
> contacted some companies and they were not interested, for example I could
> go a little far with Intel thanks to a very nice and impressive engineer
> from Intel who wrote an article on the new codecs and quoted my work, he
> passed on my work to an Intel video compression director and the answer was
> that the NHW Project is not interesting... I must also say that this video
> compression director is the Intel MPEG representative...
> So very soon, when the new video compression standard: AOM or MPEG
> establishes, the NHW Project will quickly disappear (or die of old age), so
> finally I don't know if it's really worth to make a website for the NHW
> Project whereas it will soon disappear...
> There was still the opportunity to replace MotionJPEG, but the author of
> the study told me that in fact MotionJPEG is not used for its performance
> but because JPEG is widely supported...
> So again, despite all the encouragements I had on this channel (and I
> thank you so much again), I still think the NHW Project has not and will
> not find an interest in the industry and will soon disappear with the new
> standards coming.
> For now, I am waiting for the confirmation that I will start my new job in
> September-October (alternating Master's at University and with a company),
> and I don't know if I will really have time for the NHW Project then, I
> have also furthermore an other project that interests me a lot...
> Are there motivated people who wants to take over the NHW Project from
> September-October? Spork? Others?
> Cheers,
> Raphael
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.xiph.org/pipermail/theora/attachments/20180719/091c15f4/attachment-0001.html>

More information about the theora mailing list