Talk:International Data Encryption Algorithm

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

TODO list (now done!)[edit]

It's late and I'm tired, but this article lacks information on (a) the input/output transformation stage, and (b) the subkey generator. →Raul654 03:19, Mar 21, 2004 (UTC)

True, but they aren't very interesting... Matt 03:36, 21 Mar 2004 (UTC)
I'll add them to the to-do list anyway. — Matt 23:24, 27 Nov 2004 (UTC)
I have added an image showing the transformation round. danielap48 02:04, 19 Oct 2006 (UTC)

Image format[edit]

Erk, it looks very ugly with a grey border around it; is there any reason you wanted it 200px and not 250px?

— Matt 20:47, 18 Apr 2004 (UTC)

The manual of style says not to use HTML-formatted pictures like that - it's depricated since we got the wiki image markup language. If you don't want the grey border, then simply remove the "thumbnail" switch. However, doing that removes the caption. Tell you what - read Wikipedia:Picture tutorial (I wrote it :D) and tell me what format you would prefer. →Raul654 20:51, Apr 18, 2004 (UTC)
Is there no way to have no border AND a caption? — Matt 20:53, 18 Apr 2004 (UTC)
Only with html tables, and we generally discourage that. If you do it with no thumbnailing (no grey border), then the caption shows up if you hover the mouse over the picture. →Raul654 20:55, Apr 18, 2004 (UTC)
Hmm, that's unfortunate; I've flicked between the versions, and it does look quite a bit uglier (IMO) with the thumbnail box, but really could do with the caption rather than be a bare image. Could we just leave it as is for now? I'm all in favour of removing nasty HTML stuff from wikitext, but it might still be best for treating floating white-background diagram-type things ... — Matt 21:11, 18 Apr 2004 (UTC)
It does look a lot nicer in the new skin! — Matt 11:03, 30 May 2004 (UTC)[reply]

Diagram[edit]

Can someone please edit the image to make it clear when paths cross over, perhaps with the little bridges seen on wiring diagrams? Pimantony 19:41, 27 March 2006 (UTC)[reply]

Incorrect Key Schedule description[edit]

The key schedule is incorrectly described. The 25 bit left rotate is not applied to the 128 bit key until after the full 128 bits are used in the key. This happens in the middle of the second round (96 bits are used for the first round, the remaining 32 bits are used in the second round.) The rotate left operation is only performed a total of 6 times in the key schedule. —Preceding unsigned comment added by 216.83.134.34 (talk) 03:35, 21 June 2008 (UTC)[reply]

External Link[edit]

The IDEA applet link does not appear too useful, not in English and I saw no way to run a simulation.--Billymac00 (talk) 19:24, 4 January 2009 (UTC)[reply]

patents[edit]

The article says "IDEA is patented in Austria, France, Germany...". you cannot patent software (or any other algorithms) in germany.... — Preceding unsigned comment added by 213.61.9.74 (talk) January 26, 2011

Patent expiration date[edit]

The term of patent in the United States has been questioned by Fgrieu (talk), so I'd like to collect discussion here. Because the patent was filed prior to June 8, 1995, when the patent term rules changed, its term is the longer of the two rules:

  1. (Old rule) 17 years from date of issue. May 25, 1993 + 17 years = May 25, 2010.
  2. (New rule) 20 years from date of filing. (35 U.S.C. § 154(a)) But the date of filing includes, in some cases, priority dates based on international patent applications (see Patent Cooperation Treaty).

Before Fgrieu's edits, the date of U.S. filing (1992-01-07) was used, but he rightly points out that an earlier date might actually apply. I'm just not sure which one.... —Preceding unsigned comment added by 71.41.210.146 (talk) 20:54, 24 May 2011 (UTC)[reply]

One option for the priority date is the date of issue of the Swiss patent (1990-05-18), which would place the 20-year expiration on 2010-05-18, slightly before the "old rule" date.

A second option would be the 1991-05-16 date listed on the USPTO Patent Information server (cookies & captcha required to use) under the "Foreign Priority" tab for the United States. That would place expiration on 2011-05-16, a few days ago.

It all has to do with which subsection of 35 U.S.C. § 365 applies. § 154(a)(2) says that a date from § 365(c) does apply, but § 154(a)(2) says that a date from § 365(a) or (b) does not. I don't really understand this designation thing. Does someone else? 71.41.210.146 (talk) 20:45, 24 May 2011 (UTC)[reply]

My understanding, based on talk with a real patent lawyer about another case, is that the "foreign priority" section of the USPTO records contains precisely which foreign applications have been designated by the applicant as giving the benefit of an earlier priority date, which filing date(s) serves (under the current rules) as the reference for the end of the 20 years protection period. That benefit, I was told, is avoidance of self-anteriority. Could this be an oversimplification, and could a foreign application be in "foreign priority" with no effect on the 20 years protection period? Fgrieu (talk) 05:20, 25 May 2011 (UTC)[reply]
Formally, my reading of 35 U.S.C. is that, per section 154(a)(2), the 20 years period starts on "the date on which the application for the patent was filed in the United States or, if the application contains a specific reference to an earlier filed application or applications under section 120, 121, or 365(c) of this title, from the date on which the earliest such application was filed." (emphasis added); I action the right side of the or and posit that the Swiss patent filed 1990-05-18 (and the European/WIPO patent filed 2011-05-16) falls under section 120 as it is referencing section 363: "An application for patent for an invention disclosed (..properly..) in an application previously filed (..) as provided by section 363 of this title (..reading:) An international application designating the United States shall have the effect, from its international filing date under article 11 of the treaty, of a national application for patent regularly filed in the Patent and Trademark Office except (..)". Fgrieu (talk) 05:50, 25 May 2011 (UTC)[reply]
About the restriction to "an international application designating the United States" (emphasis added): when filing an EPO or WIPO patent application, the inventor/applicant specifies a list of Member States in which application of that patent is though. I think that can includes the USA, at least for WIPO. I have no idea if that exists in Swiss patent applications. Fgrieu (talk) 06:38, 25 May 2011 (UTC)[reply]
Reading the Patent Cooperation Treaty web site, it says that an international application designates certain countries, and is a separate step after the original patent is issued. (Both U.S. and Switzerland are signatories of the treaty since 1974, BTW.) The U.S. law, and the explanation at that web site, seems to refer to the date of the PCT application, not the original patent.
Given that a PCT must be filed within 1 year of the original patent application, and starts the clock on the patent validity on other countries, it sure looks like the 1991-05-16 date is the "just under the wire" PCT filing date, which § 365 and § 363 are referring to. 71.41.210.146 (talk) 13:39, 25 May 2011 (UTC)[reply]

After some research, I think I understand it. I updated the article accordingly. Comments are welcome. 71.41.210.146 (talk) 15:38, 25 May 2011 (UTC)[reply]

I support unreservedly the conclusion that after 2011-05-16 the U.S. patent grants no protection, and I'm fine with the new status of the article. Thanks! Fgrieu (talk) 23:28, 25 May 2011 (UTC)[reply]
Further, this reference [1] implies that priority date does has influence on novelty but not protection timespan; which implies that the protection period ended 2011-05-16 for the Europe too. This is even recorded in the Swiss patent status [2]. Fgrieu (talk) 15:40, 30 May 2011 (UTC)[reply]
I think that this analysis contains way too much original research. We have to be careful not to seduce people to use IDEA as if it were in the public domain, and later it turns out that your analysis was wrong because of subtleties and people relying on Wikipedia face legal trouble. Let's keep 2011–2012 and January 7, 2012 to be on the safe side. Once there is some kind of official statement that licensing is no longer granted because of patent expiry, we can give a definitive date.--rtc (talk) 22:59, 31 May 2011 (UTC)[reply]

Avoiding the OR, this site [3] has a wonderful list of many crypto patents and their expiration. For IDEA, it says (I've added links to two of the patents that I could find):

US Patent: 5,214,703[4] expiry date 25.05.2010
Europe Patent: EP0482154[5] expiry date 16.05.2011
Japan Patent: JP322544B2 (sic, should beJP3225440B2) expiry date 16.05.2011

This is consistent with my read of the two patent docs with links above (I am a U.S. patent lawyer, but my OR is just as O). Edit: See below; the patent expired May 16, 2011 in the U.S., too. TJRC (talk) 23:29, 5 July 2011 (UTC)[reply]

Some additional sources:
Web
  • PGP FAQ: "This patent expires 25 May 2010 (USA) or 16 May 2011 (Europe and Japan)."
Printed books
  • Data privacy and security, by David Salomon, p. 178: "Furthermore, the patent (which will expire in 2011) restricts its commercial use."
  • Java I/O, by Elliotte Rusty Harold, p. 238: " The patent expires in 2010 in the US, 2011 in Europe."
  • Cryptology unlocked, by Reinhard Wobst, p. 232: "Unfortunately, it is less noticed that IDEA is patent-protected—in Europe until May 16, 2011, while the USA will enjoy its free use one year earlier." See also the table 5.1 on page 233, which lists the various IDEA patents and their expiration. Note that at the time of the books 1996 publication, the Japanese patent had not yet been issued.
  • Official (ISC)2 Guide to the CISSP CBK, Second Edition, by Harold F. Tipton, p. 349: "The patents on IDEA will expire in 2010–2011, but it is available for free for noncommercial use."
Non-RS web sites, such as forums and blogs, are mixed, with many referring to a 2012 expiration, but I suspect they get their info from this article. One site is express about that: "the European patent protecting the IDEA (International Data Encryption Algorithm) block cipher has expired a few days ago, on May 16th, 2011, and hence felt into the public domain. Note however that, according to its Wikipedia page, the cipher is still protected in the US until January 7th, 2012." TJRC (talk) 00:05, 6 July 2011 (UTC)[reply]

Explanation of U.S. patent expiration[edit]

Here's an explanation of the US patent expiration for U.S. patent 5,214,703. I acknowledge this constitutes original research; I offer it not as a source, but rather to explain why we should have confidence that the 2012 expiration date does not apply. I am a U.S. patent attorney, but why should you believe that?

Statutes and relevant (and not-so-relevant) dates[edit]

Here's what we're going to look at:

Now, here are the interesting dates to note, from the face of the patent:

  • 1990-05-18: Initial Swiss patent application no. 01690/90 is filed.
  • 1991-05-16: Patent Cooperation Treaty (PCT) application (pub. no. WO90/12452) is filed.
  • 1992-01-07: PCT application enters "national stage" in the United States
  • 1993-05-25: U.S. Patent no. 5,214,703 issues.

Analysis[edit]

The article at present states, incorrectly:

(According to the PGP FAQ[7], the US patent expired on May 25, 2010. However, US patent law was changed in 1995 such that patents now expire 20 years after filing, not 17 years after issuing. This holds retroactively for all patents that had not yet expired at the time the changed law came into effect,[8] and it thus holds for IDEA. International treaties may or may not cause expiry as early as May 16, 2011.[9])

The italicized portion is a slight oversimplification, although harmless here, since it ends up in the same place. Section 154(a) is the provision that generally covers patent term, and yes, it was changed in 1995 (actually, it was changed December 8, 1994 with the passage of the Uruguay Round Agreements Act; becoming effective for our purposes six months later, on June 8, 1995). However, it is not retroactive, as such, to pre-June 8, 1995 patents and applications. Rather, the term of those patents and applications are governed by § 154(c). That section says:

The term of a patent that is in force on or that results from an application filed before the date that is 6 months after the date of the enactment of the Uruguay Round Agreements Act [i.e., June 8, 1995] shall be the greater of the 20-year term as provided in subsection (a), or 17 years from grant, subject to any terminal disclaimers.

In other words, the 20-year-from-filing "new rule" is not applied retroactively as the article states; rather a new hybrid rule applies to these patents: they get either the term under the "old rule" (issue + 17 years) or the "new rule" (filing + 20 years), whichever is longer. So it's not simple retroactivity.

The term under the old rule would have been May 25, 1993 + 17 years = May 25, 2010. We'll see later that this is the lesser of the two and won't matter, but let's put that aside for a moment.

Now, all we have to figure out is 20 years from the filing date. But the question is: which of the three "filing dates" is used? I'll go to the statute in a moment, but the basic principle (just in case you want to make sense of this) is that the filing date that first asserts rights under U.S. patent law is the one that's used. This is true even if the prosecution of the application, where the examiner compares the claimed invention to the prior art to determine whether a patent should issue, honors an earlier filing date to determine whether art is "prior art" or not.

Section 154(a)(2) says applications filed under sections 120, 121, and 365(c) count. Section 154(a)(3) says that applications filed under sections 119, 365(a) and 365(b) do not. We'll do the 154(a)(2) set first and then the 154(a)(2) set.

Sections 120 and 121 are not at issue here. These sections cover prior U.S. applications that describe the same invention. They're essentially the same thing. Section 120 is directed at a prior application that either resulted in a patent, and then continued to be processed in a second application to get additional claims, or a prior application that did not result in a patent, and was refiled to continue prosecution. Section 121 is directed at an application that the Patent and Trademark Office (PTO) determined contained two or more distinct inventions, and so required that the application be divided into multiple applications. In either case, the idea is that the invention was described in an earlier U.S. patent application. None of the applications at issue here are 120 or 121 applications.

Section 365(c) refers to a PCT application (called an "international application" in the statutory text; that's defined in section 351) that designates the United States. That's exactly what we have here: the 1991-05-16 PCT application. We know it designated the U.S., because it ended up as a U.S. patent. So, for the moment, our working filing date is 1991-05-16. Again, the principle at work here is that this is a patent application that, if maturing to a patent, would assert U.S. rights.

We can ignore the national stage 1992-01-07 date. This is the date the PCT application turns into what is effectively a regular U.S. application. If there was no prior PCT application, just a regular patent application, this is the date that would be used; but due to 365(c), the PCT application date is used instead.

Okay, so based on 154(a)(2), 1992-01-07 is the filing date. But now let's discuss those the other sections, which 154(a)(3) say do not count: 119, 365(a) and 365(b).

Section 119 covers the case where a non-PCT application is made in a foreign country. U.S. patent law allows you to file a U.S. patent application, claiming priority from an earlier non-US application. Rather than filing a PCT application, Ascom could have filed a U.S. application, claiming priority from this Swiss application. In that case, the Swiss application would have been a § 119 application. It would have been used to determine what is and is not prior art, but would not have been used to start the patent term clock (because that application, unlike the PCT did not claim any U.S. rights; that would not have occurred until Ascom filed its U.S. application).

So the Swiss patent application is somewhat like, but not quite, a 119 application. It doesn't qualify as a 119 application, because the U.S. patent is not based on a U.S. application that claims priority to the Swiss application. Instead, the U.S. patent is based on a national-stage patent application, which derives from the PCT application, which derives from the Swiss application. In any event, even if it were a 119 application, it doesn't enter into determining the term; § 154(a)(3) says to ignore it. So, the 1990-05-18 date can be ignored.

Sections 365(a) and 365(b) are essentially the PCT equivalents of 119. 365(a) essentially covers a PCT application that designates a non-US country. This can be treated effectively like a 119 application: a regular U.S. application can be filed based on it, just as with a 119 application, and the 365(a) application will be used to determine priority date; but it will not start the 20-year term.

365(b) essentially covers a PCT application that does designate the U.S., and that relies on an earlier non-US filing date. That's what we have here. The 1991 PCT application claims priority from the 1990 Swiss application. In this case, the Swiss date will be used as the priority date, but not to compute the patent term. It's a little odd, because the PCT application comes in under both 365(b) and 365(c). What this means is that, per 365(b), the Swiss application will be used for priority, but because of 154(a)(3) will not be used to compute patent term; however, per 365(c), the PCT application itself will be used to compute the term, because of 154(a)(2).

Conclusion[edit]

So the upshot is: the filing date used to compute the term is the PCT filing date, 1991-05-16. That means the patent expires on the later of the following two dates:

  • Filing date (1991-05-16) + 20 years, or 2011-05-16
  • Issue date (1993-05-25) + 17 years, or 2010-05-25

The later of these two dates is 2011-05-16; and that's the expiration date: May 16, 2011.

I know this is kind of muddy, but I hope it cuts through some of the confusion in the discussion above, and the 2012 date does not apply. Although some of the sources say 2010 (erroneously) it's clearly not 2012. TJRC (talk) 00:00, 7 July 2011 (UTC)[reply]

The overwhelming majority of all people reading this article will do it from a very practical perspective: They want to know after what date IDEA can safely be used without the fear of patent infringement. They don't read it for theoretical purposes about what interactions of what laws might cause exceptions. And from this perspective, the article states (correctly, it may be added) that IDEA is safe to use after January 7, 2012: it is "expiring not later than January 7, 2012". It doesn't deny it may be legal to use it before that date, but this question has not been discussed before except for what has been written above. The article acknowledges that "expiry [may take place] as early as May 16, 2011", but, as you admit yourself, this "constitutes original research". You present a complex and long analysis above to support this claim. Such analyses are sometimes wrong (in less obvious places)—without any bad intent. The world is a complicated thing. Even for a U.S. patent attorney. ;) I hope we can agree that the article presents the situation in a fair manner. Repeating your arguments in the article, in addition to what is said already, would give it undue weight. --rtc (talk) 17:31, 6 August 2011 (UTC)[reply]

Given that it is now after Jan 7, 2012, is it safe to say that under any interpretation, the IDEA patent has now expired in the US? Riordanmr (talk) 18:58, 16 February 2012 (UTC)[reply]

Patent Situation in 2012[edit]

As others have noted, this article really needs to update the current patent situation. It's quite certain that the US patent has expired, but I hope someone is able to find a good source for the other patents granted. MMcCallister (talk) 16:30, 23 August 2012 (UTC)[reply]

External links modified[edit]

Hello fellow Wikipedians,

I have just modified one external link on International Data Encryption Algorithm. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 00:53, 15 November 2017 (UTC)[reply]

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:International_Data_Encryption_Algorithm&oldid=1202974341"