There’s not really a better way when you’re a monopoly. That’s the problem.
With QUIC and with webp, there was no period of time where the new protocol/format had to compete against other experimental options to see which would win out.
Because Google put it out, and they control an overwhelming share of clients and servers, they were both a foregone conclusion. Google released it, so now it’s a standard. Other companies can either adopt it or fall behind.
This allows them to stack the deck in favor of their portfolio, even if other options were technically superior.
Or I should say, the several supported horses it rode in on, and the rest of the horses standing there looking confused like they have no idea what a webp is even though they can make a preview thumbnail of the webp image, but just can’t do anything else with it.
Except webp has worse compression, doesn’t support HDR and has a max colour depth of 8bpc. What makes JPEG-XL great is that it covers loads of use cases and is very future proof.
Webp is “current-useless” on account of being unsupported by a lot of software, including Google’s own office suite. Just the same as jxl.
At the end of the day, any standard or protocol that is not widely supported is de-facto useless. Some examples are: ipv4 multicast, TCP multipath, MIR, hashcash; all of this are technically valid, but nothing supports them, so it doesn’t matter
Except .jxl is backwards compatible with .jpg. You can losslessly convert a .jpg to a smaller .jxl file. Also .jxl is future proof and supports ridiculously high resolutions up to 1 terapixel.
Yeah, I guess if we’re being pragmatic/reasonable, you are right. I think jxl use and general adoption would be much more of a symbolic win than a technical one. The problem is that a new image format only reaches general adoption if Google supports it. Webp is a decent Google-developed standard that has reached general adoption because Google pushed and supported it and if Google doesn’t support jxl, then you can’t use it on the general internet.
This is not about jxl being the better standard, this is about the better standard being non-viable because Google doesn’t want it to be.
The caveat here is that jxl is not half as old as webp and this situation might change drastically in a matter of months (I bet ff would be quick to adopt jxl if Google did…) ((although the jxl standard is technically 3 years older than the webp standard))
This might be an unpopular opinion, but webp fulfills the same use cases and some more. But it’s not jpeg and some SW still doesn’t support it.
Kinda missing the point here.
One company shouldn’t be able to pick winners and losers for file formats or protocols.
Google has done it over and over again: https://en.wikipedia.org/wiki/QUIC
Pretty sweet gig, being able to give yourself an 8-year head-start.
They released the gQUIC reference implementation under the BSD-3 licence. What would have been your preferred way?
Google is a monopoly, consequently most of their in-house tech will make waves if they decide to use it.
Or are you suggesting a nefarious purpose, like MS’s EEE?
There’s not really a better way when you’re a monopoly. That’s the problem.
With QUIC and with webp, there was no period of time where the new protocol/format had to compete against other experimental options to see which would win out.
Because Google put it out, and they control an overwhelming share of clients and servers, they were both a foregone conclusion. Google released it, so now it’s a standard. Other companies can either adopt it or fall behind.
This allows them to stack the deck in favor of their portfolio, even if other options were technically superior.
Fuck webp and the horses it rode in on.
Or I should say, the several supported horses it rode in on, and the rest of the horses standing there looking confused like they have no idea what a webp is even though they can make a preview thumbnail of the webp image, but just can’t do anything else with it.
Except webp has worse compression, doesn’t support HDR and has a max colour depth of 8bpc. What makes JPEG-XL great is that it covers loads of use cases and is very future proof.
Webp, on the other hand is “current-proof”. JXL’s biggest issue is getting adoption, not the technical aspects.
The meme is right, chrome/ium support would do a lot to help adoption. But not even the mainline FF version has support (AFAIK).
stable firefox has an option to support it, in about:config that is set to false by default
Webp is “current-useless” on account of being unsupported by a lot of software, including Google’s own office suite. Just the same as jxl.
At the end of the day, any standard or protocol that is not widely supported is de-facto useless. Some examples are: ipv4 multicast, TCP multipath, MIR, hashcash; all of this are technically valid, but nothing supports them, so it doesn’t matter
Except .jxl is backwards compatible with .jpg. You can losslessly convert a .jpg to a smaller .jxl file. Also .jxl is future proof and supports ridiculously high resolutions up to 1 terapixel.
JXL is better on paper/technical specs, but adoption is terrible while wepb is already here as it replaced gifs and most web images.
I have nothing specific against JXL, it’s a good format, but I don’t feel its necessary either.
Webp adoption also used to terrible, but now it’s mostly fine.
Yeah, I guess if we’re being pragmatic/reasonable, you are right. I think jxl use and general adoption would be much more of a symbolic win than a technical one. The problem is that a new image format only reaches general adoption if Google supports it. Webp is a decent Google-developed standard that has reached general adoption because Google pushed and supported it and if Google doesn’t support jxl, then you can’t use it on the general internet.
This is not about jxl being the better standard, this is about the better standard being non-viable because Google doesn’t want it to be.
The caveat here is that jxl is not half as old as webp and this situation might change drastically in a matter of months (I bet ff would be quick to adopt jxl if Google did…) ((although the jxl standard is technically 3 years older than the webp standard))