Article View: comp.arch
Article #12422Re: big + little endian (was: Comparison of Alpha, MIPS ..)
From: Mark Gilbert
Date: Mon, 04 Jan 1993 17:49
Date: Mon, 04 Jan 1993 17:49
51 lines
3079 bytes
3079 bytes
In article <WAYNE.92Dec30093950@backbone.uucp> wayne@backbone.uucp (Wayne Schlitt) writes: >From: wayne@backbone.uucp (Wayne Schlitt) >Subject: Re: big + little endian (was: Comparison of Alpha, MIPS ..) >Date: 30 Dec 92 15:39:50 GMT >In article <1hquujINNmov@fido.asd.sgi.com> pj@sam.wpd.sgi.com (Paul Jackson) writes: >> |> 01101001 may mean `105' (big-endian) or `150' (little-endian) >> >> Big and little endian refer to the order of bytes within words, >> not to the order of bits within bytes. >> >Big and little endian refer to the order of sub-units within units, >not just bytes and words. >You can order bits within bytes either way, bytes within words, and >words within double words. Various computers systems have made >different choices at all of those levels. The ordering of bits within >bytes has pretty much standardized on the little-endian format. On a system that addresses bytes, the bits within the byte do not have an ordering, except in a graphics sub-system, where the msb can be displayed on the left or the right. (Assuming a pixel size smaler than 8 bits.) Unfortunately, on mainstream PC graphics subsystems, the most significant bit of a byte is displayed on the left, even though the system is othewise little-endian. This forces software to perform non-trivial operations on raster images a byte at a time. If a unit larger than a byte is read, the bits within that unit are mis-ordered. I have identified the root of this whole big-endian/little-endian debate. The conflict is is between text and numbers. We read left to right. But, if you were to number the digits of an integer, you would do it right to left. The one's digit would be 0, the ten's digit, 1, etc. This should never have been the case. We adopted our number system from the Arabs, and the Arabs read right to left. For them, the direction of reading text and numbers is CONSISTENT. The whole problem arose because we failed to reverse the numbering system when we borrowed it from the Arabs. I propose that we now correct this historical oversight. Let us write numbers least significant digit first. The bang symbol can be used to replace the decimal point, in order to distinguish little-endian numbers from big-endian. The integer six hundred and fifty four should now be written !456. Sixteen dollars and twenty five cents is now $ 52!61. Sincerely, Mark Gilbert
Technical Details
Message-ID:
<markg.79.726169747@county.lmt.mn.org>Path:
arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!agate!dog.ee.lbl.gov!news!nosc!crash!orbit!lmt!ccdtc.lmt.com!markgReferences:
<1992Dec29.044012.1@cc.curtin.edu.au> <3623363@zl2tnm.gen.nz> <WAYNE.92Dec30093950@backbone.uucp>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!agate!dog.ee.lbl.gov!news!nosc!crash!orbit!lmt!ccdtc.lmt.com!markg From: markg@county.lmt.mn.org (Mark Gilbert) Newsgroups: comp.arch Subject: Re: big + little endian (was: Comparison of Alpha, MIPS ..) Message-ID: <markg.79.726169747@county.lmt.mn.org> Date: 4 Jan 93 17:49:07 GMT References: <1992Dec29.044012.1@cc.curtin.edu.au> <3623363@zl2tnm.gen.nz> <WAYNE.92Dec30093950@backbone.uucp> Sender: usenet@lmt.mn.org Organization: LaserMaster R&D, Minneapolis, MN USA Lines: 51 Nntp-Posting-Host: ccdtc.lmt.com