GIF Slips Away From Unisys; Your Move, IBM 609
Twenty years ago, Terry Welch's improvement on Lempel-Ziv compression appeared in IEEE Computer magazine. The authors of unix 'compress' and the GIF standard incorporated that algorithm without realizing it was patent-pending. When the submarine patent surfaced ten years later, its new owner Unisys intimidated developers and web authors into moving away from GIFs, inspiring the creation of a better standard, though sadly still a less popular one. Today, July 7, 2004, Unisys's last LZW patent (in Canada) expires, leaving GIF once again free... almost. See, there's the small matter of IBM's patent, granted on the same algorithm, which is valid for another two years. That still has a chilling effect on GIF development, though the consensus seems to be that IBM would lose any court action it tried to bring. So how about it, IBM? You've got nothing to lose! Want to make a lot of geeks happy and release that final patent into the public domain?
GO IBM! (Score:5, Insightful)
If the poster is correct (Score:5, Insightful)
This is cause for celebration. (Score:3, Insightful)
Anyone happen to have a copy of the alg. lying around?
Re:If the poster is correct (Score:5, Insightful)
Well, one reason is that, once-upon-a-time, we had to build apps for browsers that didn't support .png, so even though we could handle .pngs, we had to consider our clients who were stuck with .gifs. Thankfully, even the lowliest [microsoft.com] of browser almost supports .png these days.
Why should we care (Score:5, Insightful)
What would be more interesting is suing someone over it. This patent "cold war" is annoying - it would be more beneficial to see an all-out war where large companies crumble, and the idiocy of software patents is demonstrated once and for all. Cold war only server to suffocate, and masses never learn of the damage being done, because it's so invisible.
Interesting article on how IP law conflicts with ancient chinese tradition is here [slashdot.org]
Re:If the poster is correct (Score:3, Insightful)
The other patent should constitute prior art (Score:3, Insightful)
Of course, the lawyer types might still want to argue the case since that's how that make their money
Comment removed (Score:5, Insightful)
IBM is friendly (Score:1, Insightful)
Re:Why do we need GIF anymore? (Score:3, Insightful)
Chest Thumping (Score:4, Insightful)
PNG (Score:5, Insightful)
</TokenMicroSuckJab>
jamie needs to hit the books. (Score:4, Insightful)
Why yes, nothing to lose. Which is exactly why you're practically begging them.
... though the consensus seems to be that IBM would lose any court action it tried to bring.
No offense jamie, but you should really refrain from making things up like this. There is no one anywhere with any sort of legal background that would agree with this. Hell, it's probably libel to say that. It most assuredly is an outright lie.
If IBM releases it, then that's great, but don't try to badger them into it.
Re:If the poster is correct (Score:4, Insightful)
Two issues with this:
1. THe fact that you cannot imagine a practical use doesn't mean there isn't one.
2. Seeing the popularity of flash to create animated advertisements, banners and such, there may not be a 'need', but defintiely a very strong wish to use such things.
Am I missing something? (Score:5, Insightful)
I personally think having LZW is of much more significance than GIF.
Re:If the poster is correct (Score:2, Insightful)
Re:If the poster is correct (Score:1, Insightful)
Because some websites require animation, and don't want to go as far as Flash or embedding quicktime/other video files. There's more to GIFs than the dancing hamsters, dummy.
Re:If the poster is correct (Score:2, Insightful)
IBM isn't going after anybody.. (Score:4, Insightful)
Unisys was collecting money on GIF licenses for years, if IBM wanted to capitalize on this, they would've sued Unisys back then.
Besides that, there is good reason: It is, by all accounts I've read, the same algorithm.
The Unisys LZW patent had even been granted before the IBM patent had been applied. It had priority by a mile. The IBM patent is simply worthless.
Developers shouldn't concern themselves with bogus patents. I for one have written programs which save GIF files, and although I respect(ed) the Unisys patent, I'm not at all worried about the IBM one.
Re:not even close! (Score:2, Insightful)
if the browser with 95%+ market dominance doesn't support its most useful feature...
...then what's stopping you from using it? 5% or browsers will benefit; and you'll benefit from not having to update your site in 1..100 years when MS release the next IE.
Still, if you're happier using a proprietary format that may still be patent-encumbered, in preference to an open format, hey! That's your call.
Re:If the poster is correct (Score:3, Insightful)
Actually, it's JFIF -- the JPEG File Interchange Format.
Re:If the poster is correct (Score:1, Insightful)
transparancy:
- non-square logos that need to be applied to areas with different background colors.
- A logo which is randomly selected from a number of possible gifs and overlayed on top of a very complex picture. Saves space if you don't have to save the complex background picture into every logo.
- Take a look at the average discussion forum where people use avatars.
As for animations, those are used often for far more useful things than dancing hamsters and nonsense like that.
Re:PNG's..... (Score:2, Insightful)
Animated GIFs are an ugly hack that the PNG people didn't want to repeat. Splitting animated and non-animated images into two formats was the right thing to do. Unfortunately, since MNG supports so many fancy features beyond just simple animations, its adoption has been virtually nil - unlike PNG which is widely supported even on Windows with IE (for the most part).
Re:in any case (Score:5, Insightful)
Your comment makes no sense in this context.
IE doesn't support alpha transparency in PNGs, and that's substandard on their part, but I don't think the web would change much if it did unless everybody started bloating their sites with transparent effects where it is not needed.
You couldn't be more wrong. If people could use PNG the way it's supposed to be used, we could have rounded corner graphics that don't suck, change background colors without having to modify all images to match, have different background colors on different pages without the need for extra graphics for each different color background, allow user-selectable page colors, et cetera. It would actually save a lot of bandwidth.
As it is, there is very little benefit to using PNG in most cases, so people don't switch.
And PNGs with alpha-transparency are not "bloated" by any means.
Re:LZW tiff, too (Score:3, Insightful)
Either way, you're safe.
LZW is USEFUL and NON-OBVIOUS (Score:5, Insightful)
It was a big breakthrough when algorithms like LZW, which compressed data that contained repeated multi-byte patterns (like text, or bitmap drawings), were developed. The previous state of the art was to pre-analyze the data and build a table that would have to be exchanged before the data could be decompressed (like Huffman encoding). LZW lets you built the table on-the-fly as the data is compressed, and exchange it on-the-fly as its being decoded (because the compression "table" and the data stream are actually the same.)
LZW does seem simple to us now; in fact one standard Job Interview question I ask is to put the LZW algorithm on the whiteboard! However, for those of use who have been around for more than 20 years, it was a significant breakthrough.
hmm. (Score:3, Insightful)
A quick cursory overview of the patent link on IBM's patent doesn't say one thing about the GIF format, just the compression algorithm (with JCL code).
What if this patent doesn't cover GIF at all, but a hardware implementation of compression on a hard drive, or a MO drive, or some other device? They can't exactly release all claims to it that easily.
Just seems silly to 'call out' a company to release a patent. Contrary to popular belief the bigger companies out there can't turn on a dime and have hundreds of processes to do things to keep a rogue employee from releasing all claim to all patents or something crazy like that, so it could take them two years just to release something that's going to die quietly anyway.
Also speculating on what a company will/won't do with a patent based on some arbitrary IANAL comment from the editor seems a bit risky. While IBM is into Open Source heavily they're not there to stop making their stockholders money either. Patenting things lets them do so.
Re:If the poster is correct (Score:2, Insightful)
Re:not even close! (Score:5, Insightful)
Some people don't have the option to live in an ideological world and must live in a realistic one.
Elsewehere in this thread I've mentioned that I continue to use .gifs, because it's not realistic to ditch them. Regardless; I'd hardly descibe my position as "ideological" or "unrealistic" - it's one based on:
1. Portable Network Graphics work with all major browsers - now (IE doesn't support one area of PNGs, but it doesn't lose any functionality over GIFs, as GIFs don't provide 8-bit alpha-blending anyway);
2. GIFs may - as this article is about - still be patent-encumbered.
In what way is promoting increased use of PNGs unrealistic?
Intimidated? (Score:3, Insightful)
Not very well I hasten to add, GIF's are still used rather a lot [google.com] and even Slashdot hasn't bothered to convert all their images to PNG.
Re:jamie needs to hit the books. (Score:3, Insightful)
He's free to say that there is a concensus without providing links. The consensus could well have been between his two lawyer friends.
Animations... (Score:3, Insightful)
Native support in the browser for SVG and SVG animation would more than replace animated GIFs as well as providing lots of interesting capabilities that could be useful in other areas.
Of course, that too would be left unused because IE doesn't support it (or worse yet IE would support some bizarre proprietary MS reworking of the basic ideas).
oh well.... (Score:3, Insightful)
You see, that was "the real world" back then, the "practical" world. It was "impractical" to go against thw societal norm, and in that case it was physically impractical, as you could have been gassed, beat, arrested, serve jail time, and etc. So heaven forbid you have some internet surfer be inconvenienced by a semi non standard format on your web page,you or your corporation might suffer some "inconvenience" in your profits or something. Your profits are obviously of more worth to you, so go ahead, protect your profits, that is your right.
Let's always leave things exactly as they are now, let's none of us ever go against the norm, it is impractical, we might lose money,and as we all know, money is the most important thing in the known universe,95% of the people agree, nothing is as important as money, all other aspects of society should revolve around money, it's accumulation and restriction in as many diverse ways as can be imagined. Let's all "work" for a small number of large corporations, always seek to do those things that are dictated to us by our "betters" in those corporations and pseudo legitimate governmental agencies, because they, having the most money,and the most "practical" power and influence currently, must always surely know a better way to do anything, correct? I mean, they are the majority, so they must be "correct".
Re:not even close! (Score:4, Insightful)
you were implying that we should go ahead and use the uber-PNG features, and then wait for IE to catch up to see them.
Indeed, and why not? An image is just an image in IE; who cares if only 5%, 6%, 7% (and rising) of the browsing population see it as it was intended? It'll be visible in all its glory once IE 7 is available, or once punters start making the switch to more modern browsers. In the meantime IE users won't be adversely affected, and it might even prompt them to drag themselves into the 21st century, browser-wise.
Re:jamie needs to hit the books. (Score:5, Insightful)
I traded email with several people who know the history of this algorithm and its patents fairly well. I wasn't able to get a quote from a legal expert backing this up by press time, but it hardly matters because this opinion indeed is the consensus of those I have talked to. And I mentioned the duplicate-patent issue to an IBM PR rep, who had plenty of time but didn't offer a correction.
I stand by what I wrote.
Re:If the poster is correct (Score:3, Insightful)
Yes, that is indeed what I am saying.
We can argue a lot about improper advertisement and such, but meanwhile advertisements are paying for the fact that we can discuss here at all.
So no, I didn't ignore your question I think.
Re:not even close! (Score:5, Insightful)
Switching to PNG because 5% or browsers will benefit is the unrealistic part.
Well, all browsers, including IE, can use PNGs. I presume you're refering to alpha-blending, which IE can't - currently - use? Use of non-IE browsers is steadily rising, as support for IE seems either non-existent or focused on "IE7" or whatever it'll be called. This suggusts - to me - two things:
1. Increasing numbers of people are using browsers - now - that do support alpha-blending;
2. The next iteration of IE is likely to support alpha-blending and is probably due soon (maybe soon like Longhorn, but soon, anyway...);
Either way, catering for the present-and-the-not-too-distant future doesn't seem that unrealistic.
Re:not even close! (Score:3, Insightful)
If _your_country_ is patent encumbered, I don't see it as a problem with the technology, I see it as a problem with your country.
Re:If the poster is correct (Score:3, Insightful)
You can only claim that it is "superior" in that it executes more quickly. However, it produces significantly larger files. PNG should be redefined to also use BZIP2 compression, since that produces even smaller image files than GZIP compression. (BZIP2 is also particularly good at XML data--if you ever hear anyone talking about proprietary XML compression, mention that BZIP2 is the generic method to beat, not GZIP.)
Re:If the poster is correct (Score:2, Insightful)
Re:Why do we need GIF anymore? (Score:5, Insightful)
The quality of a web site is determined more by it's substance than by it's appearance.
Good web site design doesn't even require images.
Alpha blending is not critical.
It's nice, but IMO it's ranks below "spell checker" in the hierarchy of good web site design tools.
-- less is better.
Re:not even close! (Score:3, Insightful)
Indeed they do. (Score:5, Insightful)
SCO's answer to IBM's counterclaims accuses it, among other things, of selectively enforcing it. I'm not quite sure what basis there is in law for using that as a defense, however, or if that was just boilerplate text in SCO's reply.
Re:not even close! (Score:1, Insightful)
So you don't use PNG because its translucency isn't available on internet explorer?
What do you use instead? GIFs don't support translucency at all in internet explorer, or any other browser. JPEGs? don't support translucency at all in internet explorer, or any other browser. BMP? (don't laugh, I've seen them on IE-users' websites) Flash? Do you just give people a PDF to print out on acetate and hold up over their screen?
If you're using indexed transparency, then PNG is still better than GIF (more colours available, smaller filesize for same image, better support in graphics programs, etc. etc.) and internet explorer displays indexed transparency the same whether it's GIF or PNG.
(Don't reply saying some ancient pos like IE3 doesn't support PNG transparency, you're arguing on the "95% of users", which means you're taking about internet explorer 6 ONLY)
If you're not using transparency at all, then PNG is still better than GIF. (more colours available, smaller filesize for same image, better support in graphics programs, etc. etc.)
Re:Slashdot uses GIFs (Score:2, Insightful)
Well, I can't speak for Slashdot or anyone else. But , in my experience, PNG is a great format suffering from a few problems not of the format's making.
There is the infamous lack of alpha support in MSIE. That is fine. You can use 256 colors with 1 binary transparency color. Except, there is also the gamma correction problem.
Gammas on different computers being different, the gamma correction in PNG is supposed to take care of inconsistencies, providing a correctly displayed image across the various displays. However, this is not the case. Because the gamma correction relies on correct interpretation of the gamma value (embedded in the PNG) by the display system in order to make the correction. Effectively, this makes it a suggestion, not the law. As we all know, this opens up a can of worms. As one might expect, some display systems handle the gamma correction correctly, most do not. The net result is that you get slight intensity differences in the color matching of the background and the 'transparent' color of your 256-color PNG. This makes a shoddy-looking block around your image. This is probably undesireable.
GIF has no such correction scheme, but as such, triggers no mishandling of a gamma value. Instead, most display systems match the transparent value to the nearest color on the current palette. The result is (sometimes) a slightly off-color GIF. But, this is not a problem, since people detect differences in intensity of light, not subtle color change. The alterations go unnoticed. In Web browsers, the CSS color palette is (most often) used to do this color alteration. Hence, you get a GIF whose transparent color precisely matches the background it appears over. This is an advantage.
For fairness in our consideration, there is still the fact that many times PNG can get smaller file sizes than GIF. But, there are other times GIF beats PNG. It's all a matter of the number of colors saved in the palette of the image and the level of dithering used to maintain the desired clarity of the image (because many dithering algorithms increase the amount of data saved per row of pixles in the bitmap, due to the way areas of like color are compressed).
PNG vs. JPEG
Now, there are many times PNG is smaller than JPEG. This is the case when large areas of the image are close in color and intensity. The bulk of the size in the PNG is reserved for the detailed areas. Given that PNG can be lossless, you can maintain a higher-quality image at a given file size than JPEG in these circumstances.
However, JPEG excells in details.. and will usually beat a PNG for 24 bit images with low contrast or large areas of precision detail. JPEG can simply handle detail better, since PNG gains its greatest efficiency in storing and regenerating larger areas of like color and intensity.
CONCLUSIONS
So it comes down to this:
* If you want transparent images for the Web that work in all browsers: GIF
* If you want non-transparent images for the Web with large areas of like color or intensity and that work in all browsers: PNG
* If you want non-transparent images for the Web with large amounts of detail and that work in all browsers: test between PNG and JPEG
* If you want alpha-transparent images for the Web that work in all browsers: wait for hell to freeze over
ONE MORE THING...
Oh, yeah. As for which format one should use -- its a matter of pragmatism. You may well decide that this is a FOSS vs. EvilEmpire thing, but most cannot afford the luxury of basing all of our decisions on politics/philosophy.
In a slightly more perfect world where the formats are unencumbered by implementation and outside factors, many more people may well have been inclined to choose based on our personal convictions. The best one can hope for is that we put our convictions front and center when we do have a choice. I would like to think that when a JPEG and PNG are similar in file size and image quality FOSS supporters and other idealists would use the PNG on principle.
OBLIGATORY WORDS OF WISDOM(?)
Remember, nothing casts a greater shadow upon principle than pragmatism.
Best Regards
This is why gzip exists (Score:3, Insightful)
are often in the UNIX compress *.tar.Z format.
People at the Free Software Foundation saw this
to be a problem for their GNU system project,
so they had the patent-free gzip program written.
The *BSD projects essentially beg for a lawsuit.
BTW, the original bzip (not bzip2) and the better
type of JPEG compression both infringe on a
different IBM patent. That's the one we should
want to have opened. IBM would even gain some
licensing fees if they did a GPL-only license
for it.
Transparency not a *required* part of PNG (Score:5, Insightful)
Wrong.
That's a techie urban-legend. The truth is that IE6 does support all required PNG features. Therefore it "supports PNG".
Yes, IE6 doesn't support PNG transparency, at least not in any easy way. However PNG transparency is an optional part of the PNG spec. That IE6 doesn't support transparency properly is unfortunate but doesn't invalidate their meeting the required PNG spec.
Furthermore as others have pointed out there are indeed work-arounds (ugly ones) that will enable reliable PNG transparency on IE6. Also as others have pointed out (including MS staffers) even if IE7 were to ship tomorrow and support PNG et al we'd still be stuck with a huge IE6-using population for years to come.
It would be great if IE, and indeed all of the browsers, were to fully meet all relevant standards. It would also be great if they were to then go on and meet more of the optional parts of those standards, including PNG transparency. However lets hold everyone's feet to the fire on these, not pick on one author's neglecting a feature many would like while they and others are still missing more fundamental required parts of specs.
Re:Why do we need GIF anymore? (Score:1, Insightful)
If Microsoft were to assign some engineers to IE, I'd take Web Forms 2.0 [whatwg.org] over graphics tricks. The web is pretty enough, make it more useful.
IE only has an 80% market share -- and shrinking. (Score:4, Insightful)
The 95% figure may be the Windows share of the market (more like 94.5% by that link), but not everyone using Windows uses IE. (If I'm setting up a desktop that has to have Windows, Mozilla is the first app I load on it, and then remove the IE icon from the desktop.)
The recent notices from Homeland Security about IE being unsafe will only accelerate this.
Re:If the poster is correct (Score:1, Insightful)
Anyway, my point is: "jif" is more correct, but who gives a fuck how you say it?
Re:No process within IBM (Score:3, Insightful)
I have heard that IBM uses its large warehouse of patents purely as a method of protecting itself from lawsuits. Upon digging around a bit, I found that IBM had done the same thing to Sun that it is now trying against SCO.
http://www.forbes.com/asap/2002/0624/044.html
The number of patents that IBM has surely must number in the millions by now. I wouldn't be surprised if they had patented parts of most of what you see in the computer world these days. Either that or have patents on its prior art.
I find it unlikely that IBM would really want to go to the effort of launching lawsuits against nearly everybody and everything that ever touched code and become another SCO.
We've also already heard that IBM does not prosecute their patents against open source developers.
http://bca.cryst.bbk.ac.uk/bca/cnews/2003/MAR03
So I really have no problem with them sitting on it and suing other companies that try and leverage it as intellectual property against them.
PNG Standard Hypocrisy (Score:3, Insightful)
Not that I'm trying to be flamebait for OSDN, but the very icons applied to this article (in the upper-right corner) are none other than the very GIF standard that was put down in this article. Just thought I'd point that out.
Re:Why do we need GIF anymore? (Score:4, Insightful)
It's like CSS for positioning: you don't need it (you could use tables instead), but it makes for better design, because you can separate content from presentation. PNG is like this because then the image is "my_logo.png" rather than "my_logo_over_white_background.gif", "my_logo_over_black_background.gif", "my_logo_over_blue_background.gif", "my_logo_over_puke_green_background.gif",
"my_lo
Re:SCO v IBM - This Patent Is Being Actively Used (Score:3, Insightful)
It might be strategic to drop that anyway, since it will be just about the only thing that SCO will be able to crow about winning when the case is over. Pushing an obviously bogus patent also brings suspicion on the credibility of all their other patents.
Re:Transparency not a *required* part of PNG (Score:1, Insightful)
Just implement transparency in png and put a reference to a properly working web browser like Firefox - and say "if this page doesn't work properly, then use a real web browser".
I'm sick of the "I've got a PHD and i'm realy smart" attitude from Microsoft - and I can't wait for the winds of change.
Don't let Microsoft dictate low standards to us all. This is the best form of advertising we (the people) have.
AC