Talk:SHA-1: Difference between revisions

From Rosetta Code
Content added Content deleted
No edit summary
(→‎Warning highlighting: The text is probably fine it just needs to look different)
Line 10: Line 10:


: What kind of warning is appropriate for SHA-1? [http://docs.factorcode.org/ Factor's documentation] for [http://docs.factorcode.org/content/article-checksums.sha.html SHA] says, "SHA-1 is considered insecure, while SHA-2 It is generally considered to be pretty strong." OpenBSD [http://www.openbsd.org/cgi-bin/man.cgi?query=sha1&apropos=0&sektion=1&manpath=OpenBSD+5.0&arch=i386&format=html sha(1)] gives no warning at all, though OpenBSD [http://www.openbsd.org/cgi-bin/man.cgi?query=md5&apropos=0&sektion=1&manpath=OpenBSD+5.0&arch=i386&format=html md5(1)] gives a warning and recommends sha256. --[[User:Kernigh|Kernigh]] 19:23, 19 January 2012 (UTC)
: What kind of warning is appropriate for SHA-1? [http://docs.factorcode.org/ Factor's documentation] for [http://docs.factorcode.org/content/article-checksums.sha.html SHA] says, "SHA-1 is considered insecure, while SHA-2 It is generally considered to be pretty strong." OpenBSD [http://www.openbsd.org/cgi-bin/man.cgi?query=sha1&apropos=0&sektion=1&manpath=OpenBSD+5.0&arch=i386&format=html sha(1)] gives no warning at all, though OpenBSD [http://www.openbsd.org/cgi-bin/man.cgi?query=md5&apropos=0&sektion=1&manpath=OpenBSD+5.0&arch=i386&format=html md5(1)] gives a warning and recommends sha256. --[[User:Kernigh|Kernigh]] 19:23, 19 January 2012 (UTC)
::I don't think the issue is with the content of the warning. The issue is with how it looks. --[[User:Mwn3d|Mwn3d]] 20:21, 19 January 2012 (UTC)

Revision as of 20:21, 19 January 2012

Implementations separated

We might want to set this up like MD5 and MD5/Implementation just to keep things consistent. --Mwn3d 21:40, 18 January 2012 (UTC)

Some implementations of MD5 are still on MD5 and not on MD5/Implementation. If someone creates SHA-1/Implementation, please move the Ruby implementation from here to there. If I later create RIPEMD-160, SHA-256, or so on, I might not immediately create /Implementation pages. --Kernigh 19:23, 19 January 2012 (UTC)

Warning highlighting

That highlighted warning should probably go into a centered infobox to make it more visible. –Donal Fellows 10:45, 19 January 2012 (UTC)

What kind of warning is appropriate for SHA-1? Factor's documentation for SHA says, "SHA-1 is considered insecure, while SHA-2 It is generally considered to be pretty strong." OpenBSD sha(1) gives no warning at all, though OpenBSD md5(1) gives a warning and recommends sha256. --Kernigh 19:23, 19 January 2012 (UTC)
I don't think the issue is with the content of the warning. The issue is with how it looks. --Mwn3d 20:21, 19 January 2012 (UTC)