Hey Ho, GIFs Must Go (1999)
theatlantic.com
theatlantic.com
The age of early hypertext was awash in such cases. The University of Minnesota only mentioned that they might introduce a charge for gopher servers, effectively killing the protocol. Various early conferencing and hypertext systems, including one from Sun Microsystems (NEWS?), and a few models from Microsoft, existed. Somewhere, Tim O'Reilly has commented about the early days of O'Reilly & Associates where he was looking at various options and made the early decision to not get locked into any one proprietary system, a choice which has paid handsomly.
The whole phenomenon of prortocol creation, promotion, evolution, and quite often, stagnation, is a fascinating and understudied one IMO. The role of non-market players, particularly government standards organisations, the military, and academic fields, is extraordinarily underappreciated.
Not sure I've seen protocol discussions there (just never bothered to look) but it shows a whole lot of academics and government employees in the mix, and some commercial actors as well.
It's disappointing since those programs succeeded beyond the dreams of avarice but were not designed to gather publicity. The fact that so much of our economic growth is undergirded by public research funding is forgotten even by direct beneficiaries.
There's an interesting transition as those groups moved from small (able to fit into a VW Bug, bus, or small bar) to large (auditorium, or multi-campus groups), and the relative frictions in new developments.
Past a certain scale, the easiest way to change a standard is to move forward, either abandoning or ignoring backwards compatibility.
Both software languages and various social networking protocols (there are ... many) are interesting cases in point. Mastodon is in many ways IRC v. 2.0 (and there are any number of competing contenders). There actually was (and/or is, depending on your judging criteria) a Usenet II, though it's seen limited popularity.
Another good source on much of this would be the late Pieter Hintjins. See for example:
https://mixitconf.org/en/blog/messaging---social-architectur...
Which is why I'm devastated how much new conversation about how things are put together is generally not hosted by the organizations involved, but is stored places like issues on Google Code (all gone), SourceForge (increasingly difficult and likely will begone), GitHub (which people don't seem to like to believe will one day delete their data); imagine if some of these early standardization efforts had done all of their communication on Internet protocols using the contemporary equivalent of Yahoo Groups or Google+--maybe using CompuServe or AOL?--and all of that work was now only around due to haphazard archiving efforts. (Google+ really burns me, as there was so much discussion of how things in the Android ecosystem works--particularly among all of the third-party forks, such as CyanogenMod--which is now just deleted and gone :/.)
Though given how many people have old Usenet postings they'd like to forget, maybe it's not that bad.
A history of the Brooklyn Bridge and its builder notes, apparently as a form of admiration, that he designed every detail down to the nuts and bolts. Thing is, that's because he had to. There were no standard fasteners. (The book doesn't detail this, it's just a passing reference I happened to catch.) David McCullough's The Great Bridge: https://www.worldcat.org/title/great-bridge-the-epic-story-o...
Herbert Hoover, as US Secretary of Commerce, instituted the National Bureau of Standards (now the NTIS, I think), which among other things standardised weights, measures, and yes, fasteners. Rather hagiographically addressed here: http://lcweb2.loc.gov:8081/ammem/amrlhtml/inhoover.html
Joanne Yates has written on such thrilling topics as the evolution of the interoffice memo and standarisation of business communications (and though I'm writing tongue in cheek, I really do find the work fascinating), and has been working on a book on standards and their evolution. I've not read it but it may be of interest. Engineering rules : global standard setting since 1880 https://www.worldcat.org/title/engineering-rules-global-stan...
Simon Winchester has a book on machining and measurement, also not read, though based on his previous works it should be excellent. The Perfectionists: https://www.worldcat.org/title/perfectionists-how-precision-...
There was a fairly popular book on containerised freight shipping that came out within the past five years or so called The Box That Changed the World which looks at the standardisation of containerisation: https://www.worldcat.org/title/box-that-changed-the-world-fi...
Somewhere someone commented that one of the computer containerisation standards was simply a bunch of people coming to an agreement as to how to organise and and do things. My history with Linux has taught me that various packaging formats and systems (and the infrastructure surrounding them) have profound consequences and impacts, for users, adminstrators, and software producers.
I wish I could point to a single source and say "read this", but I'm really not aware of one presently.
The details of printing signatures and folios, and having the pages come out the way you want them to when bound, is another minor fascination. As are canons of page design.
Is it? Gifs have been wildly successful. Today, many web deverlopers prefer newer formats like png for both technical and philosophical reasons, but gifs continue to make up a huge fraction of internet pictures. The average 12 year old knows what a gif is.
Depends on what you mean by "know". If you mean "moving picture that is not a video", sure they do. But if you showed them an APNG they will probably call it a GIF as well.
And simply using or supporting gifs became an issue -- it was political. That's not something you want to have happen to your standard.
This very rarely affects anything other than animations. People simply aren't using gifs for images as a matter of course today. For photos, it's largely jpeg (vastly higher compression, though lossy), for text-based images, png, though SVG is making inroads, particularly for data visualisations.
That being said, PNG always had much stronger compression, in addition to supporting 24- and 32-bit color. Nowadays, we even have WebP that usually compresses better than PNG, and compresses far, far better than animated GIF. Animated GIF is a pretty ugly hack to the format, very inefficient for many common kinds of animations (or full-blown video as it's often used now); I've found that animated WebP can produce files around 10% the size of GIFs.
I am absolutely incredulous that people choose to use a GIF over a proper video format to share video on places like Reddit in this day and age. I don't understand the motivation to do so.
Also, where have you been able to casually upload video to share quick bites like a cat making a funny face? You aren't going to upload a 12 second silent video to youtube. Up until very recently, gifs were the only permissible way to share this sort of content.
Also probably more addressable / available tools for anim gif creation.
Don't blame a web video format/standard for UI failures.
I think you don't understand the modern web. If something is crappy the user shall have it. JS was always considered crappy but now is the best new thing. And you also missed tbe latest web/UI/UX development. From the web designer point of view the user is an idiot which must be punished for every attempt to customize something.
> Don't blame a web video format/standard for UI failures.
They are not UI failures. They are done on purpose.
Video, until maybe the last 5 years, didn't.
1. Allow animations in the "Hero" user profile image.
2. Make that image permanently fixed to the top of the screen on the profile page.
3. As users increased window size to see more text and less hero, the image would scale to as much as 3/4 of the screen.
4. The hero was included on the "hover card" for users and would show up in all kinds of locations throughout the service. I still have that collection....
I spent an hour or so researching "annoying gifs", and settled on a rapid red-green-blue stroboscopic rotation. I don't know the cycle time, but it was probably a few times per second.
Google eventually relented. Well, they eventually shitcanned the whole site, but that's another story.
Trying to find animated images that weren't mind-bendingly annoying to me is how I ended up with the "space alien cat" image and avatar. The original had a rotating set of astronomical images in the glasses, I eventually applied the Crab Nebula as a static fill (tried the Cat's Eye, but ironically, it didn't work particularly well). That's the hero I use at several sites (Ello, Mastodon), and is the inspiration to the avatar I use as well.
Later at another site at which gratuitous annoying gif use was prevalent, I found a few static-but-apparently-moving optical illusion images, again to make the point that annoying images are annoying. The point was somehow missed by that crowd. That site also died.
TL;DR: animated gifs and even static images can be plenty annoying.
There is a trick where subsequent frames use transparency for non-changed parts of the image which gives a much better compression ratio than you might think.
This is so horrendously inefficient it should be a crime. But it's still a GIF, and anything that can display animated GIFs can display it too.
In other words, similar to what video codecs do with "skipped macroblocks", except at a far more granular level. One wonders how efficient it would be if GIF was extended with things like motion vectors... but between the transparency/delta-enocding and the multi-palette "layering", you can sort of see the beginnings of a video codec.
It could have very easily been better. I'm not sure what the moral is.
There’s also a lossless webp variant.
(This kind of lossy encoding would be possible with PNG too, but most PNG encoders don't have a lossy mode.)
(GIF is now patent-free, but PNG and FLIF are still better compression than GIF anyways.)
Pngs get used for both for non photographic images and photographic images with alpha channel, and JPEG XL should win on both of those use cases.
It also cleverly has an upgrade story forexisting JPEGs so I imagine a fairly quick uptake where a proxy can recompress an existing sites images and get better size with no recompression loss.
Something nice about FLIF is partial decoding. You can reasonably see a thumbnail by downloading just 5% of the full lossless file.
In conjunction with HTTP2 to progressively load all visible images first, this is a killer feature for the web, simplifying thumbnails and responsive design while vastly speeding up page loads and improving caching.
One thing that would be nice would be a way to tell whether a WebP is lossy (in which case you should only edit/resize the original file, to prevent degradation), or lossless (edit as much as you please).
In economics, the Jevons paradox (/ˈdʒɛvənz/; sometimes Jevons effect) occurs when technological progress or government policy increases the efficiency with which a resource is used (reducing the amount necessary for any one use), but the rate of consumption of that resource rises due to increasing demand.
[1] https://discussions.apple.com/thread/250073758
[2] https://developer.microsoft.com/en-us/microsoft-edge/platfor...
It's not that a bad deal though, one minute of video is roughly 7Mo for something close to DVD quality. I had GIFS lasting a few seconds weighting way more, with terrible quality.
I'm not that well-versed in the matter, is WebP already supported widely? I mean, not just on desktop and advanced smartphones?
And as a side note: png does not strictly compress better; sometimes the ability to use palette indexes that are smaller than 8 bits is an advantage.
PNG does support 4-, 2-, and 1-bit indexed modes :)
Though that's got nothing to do with why GIFs got so popular. That's purely because GIF is the only format that could display animations on every browser since 1995. The timing of PNG was rather unfortunate - the PNG team had decided on not supporting animation because GIF animation was rare and adding it to the spec would've unnecessarily complicated things. They froze the spec, planning to submit it to W3C and IETF for standardization - and then Netscape, in full "move fast and break things" mode, released Navigator 2 with animated GIF support. So soon every GeoCities page is full of annoying sparkly GIFs, PNG only lets you make boring stills, and we've got to support both formats forever.
More like a paragraph... https://en.wikipedia.org/wiki/Lempel%E2%80%93Ziv%E2%80%93Wel...
Personally, I like LZ more than LZW because it doesn't have to slowly "build up" the dictionary in the same way as LZW nor need bit-I/O (because you can have an 8, 16, or 32-bit "indicator", followed by that many literal or length/distance pairs) and thus makes for an even simpler (and faster) algorithm. The only downside is that it's more asymmetric, in that compression is usually far slower than decompression (which can be faster than memcpy() - see http://www.oldskool.org/pc/lz4_8088 ) because the former has to do a lot of searching while the latter is mainly copying memory.
For nicely optimized GIFs like these https://iwdrm.tumblr.com/ then WebP really struggles. Try to convert some of them to WebP and you'll see the result is often bigger than the original GIF.
Input #0, gif, from 'tumblr_nxt634tytr1qe0eclo3_r1_500.gif':
Duration: 00:00:08.64, start: 0.000000, bitrate: 1434 kb/s
Stream #0:0: Video: gif, bgra, 500x269, 16.67 fps, 16.67 tbr, 100 tbn, 100 tbc
Input #0, matroska,webm, from 'bar.webm':
Duration: 00:00:08.64, start: 0.000000, bitrate: 252 kb/s
Stream #0:0: Video: vp9 (Profile 0), yuv420p(tv), 500x269, SAR 1:1 DAR 500:269, 16.67 fps, 16.67 tbr, 1k tbn, 1k tbc (default)
Information is so incredibly cheap to distribute and copy today that I forget it was still very cheap to do so even before the Internet.
The dancing star gif posted in that github issue is 5,464 bytes in the original form and only 4,015 bytes when converted to a VP9 webm (with alpha channel) using ffmpeg's default settings.
> "There should always be an option to not convert to video."
I disagree. The conversion should be automatic. The converted result should then be compared against the original, and whichever is smaller should be sent to users. There may well be gifs that are smaller than their conversions, but I don't trust users to know which those are. Let the computer compare the two and make an objective decision.
It's more of a human-centric label for "short+looping+silent+autoplay video" than a tech-centric file format. A tad ironic given there's no longer a patent issue on the actual format.
https://help.imgur.com/hc/en-us/articles/208606616-What-is-G...
Nothing on imgur.com's homepage is a gif, for example. Same with gfycat.com.
First of all, it's external, not internal. There's no such thing as gifv format. It's just <video> with webm and mp4 but silent. That's the whole point. It's the whole reason why I commented and why they felt the need to create "gifv" which doesn't actually exist, lol.
How many of the things have been granted? How many filed? How much are the filing fees? How much are typical lawyering fees for it.
There are many among us who feel it is the actions of pure parasites on genuine tech innovation in software and systems so it's a question of how much was gouged by that decision (intentionally or otherwise).
The other one that could be interesting to know is the descriptive stats on salaries for patent attourneys specialising in software vs people being paid to program software.
https://bugs.chromium.org/p/chromium/issues/detail?id=514102
Animations would animate, but only play through once.
https://web.archive.org/web/20040116134137/linuxmafia.com/fa...
What's the difference (to the user) between an animated <canvas> and a silent video ?
Of course, we might be able to find a middle ground between disabling JS entirely and auto-playing all media, but it may be very hard, and that "middle ground" might be changing as website developers find new way to deliver auto-playing media that bypasses the browser's restrictions.
i've written a couple libs that help record canvas pixel art to gifs.
https://github.com/leeoniya/RgbQuant.js/
https://github.com/leeoniya/GIFter.js
btw, always compress with https://www.lcdf.org/gifsicle/ it's amazing.
It's the only format that works pretty much everywhere on every platform
We have engineers replacing tech that works with their half baked barley functioning versions that lose all the benefits of sharing and remixability of a GIF file.
Video files are easier to share, on account of being smaller, and easier to edit since you can use real video editing tools which is a lot easier than editing a gif frame by frame in gimp.
The only sense in which gifs are superior is they're easier to write a parser/renderer for from scratch, but I think that's relevant to virtually nobody.
Video: loads video player controls (slowly); waits; doesn’t start playing; starts playing but doesn’t let me choose volume quickly enough; stops playing; loading; jumps several seconds; appears to be at the end but isn’t finished loading; won’t loop; try to rewind or jump to position; doesn’t work; UI controls appear to have disconnected from anything; can’t rewind replay or watch; get annoyed; reload page entirely or quit it.
This is where you dismiss everything I’ve written with an “it doesn’t happen” because “you’re sure” videos are “perfectly playable” and “obviously work”, but you’re wrong, they don’t, they have glaring huge UX failure modes I hit all the time for years on mobile Safari with a normal person not mega fast internet connection.
Since your so skeptical, let's put some numbers to it: Here is a gif: https://0x0.st/zY8f.gif It's a staggering[0] 3.5MB, on a 56kbps connection it would take you over eight minutes to download it. I uploaded this gif to imgur, who then converted it into this mp4: https://0x0.st/zY8O.mp4 It's less than 700KB. On a 56kbps connection it would take you less than two minutes to download this. This is not an atypical example, an order of magnitude improvement is pretty typical when converting gif "videos" into real video formats.
Eight minutes vs two minutes... you tell me, which is faster? The faster your internet connection gets the narrower the gap gets, which is opposite of the trend you're suggesting. These conversions benefit people with slow internet more than they benefit those with fast internet.
[0] (For shits and giggles, I also encoded a VP9 webm with ffmpeg: https://0x0.st/zY8v.webm From what I understand you cannot play this on your iphone (because Apple), but it's less than a tenth the size of the original gif and looks identical to my eye.)
Re-read my comment this was exactly my point. We swapped something that worked everywhere for a custom players on every site and re-encoding therefore destroying what the artist intended.
No technical committee or even loose-consensus-running-code working group would ever come up with anything like that. The mode of pop-culture remixing that GIFs epitomize could never have been anticipated; the value that defaulting to silence delivers has still not gotten through to news sites (good ones, too!) that autoplay talking video over your headphones or god forbid the open plan office; it's even been reported[0] that porn works better for women in GIF mode.
That's because GIFs were never conceived for video. Consumer computer video was never conceived for pop culture remixing (despite having the experience of turntables to pick from). Porn was never conceived for women, for that matter.
GIFs are legacy left for us from a time between the fall of the Berlin Wall and the erection of the EU castle when everyone expected planning and bureaucracy to lose out to spontaneity bordering on anarchy. Many important HN recurring themes are here: privacy vs. normalized surveillance, free speech x hate speech x rogue counter-establishment right-wing speech, the freaking Mac App Store. Every "hacker" that defaults to the safe repressive position is someone who would rather that GIF looked that those barely-working WebM or whatever embeds found on Wikipedia, with little transport controls, bureaucratic efficiency and that wrongthink GIF memes were somehow blocked by "good DRM, DRM that prevents Putin from attacking the system".
Which system? The one that brought us Medium.com when we used to have Geocities. But we'll always have GIFs.
All this pop culture stuff comes over a decade from when the GIF was invented.
They do for me, because I view them with software that chooses to support such things. It really has nothing to do with the format itself.