[Home] [Products] [HowItWorks]

Gene Milligan's History of CAM ATA

Note: Gene Milligan died in November, 2000. Anyone using a data storage device owes much to Gene's efforts to develop and maintain the ANSI SCSI and ANSI ATA standards over the last 20 years. Gene will be missed as chairman of ANSI NCITS T13 and even more so as a friend.

Who is Gene Milligan? Gene Milligan has been a member of the ANSI standards committees for SCSI and ATA for many many years. Mr. Milligan is currently the chairperson of ANSI NCITS T13, the ATA/ATAPI standards committee. Back in 1989 he was the editor of the original CAM ATA document, the first public document that described what we now call the ATA interface. Several years ago he posted this history of CAM ATA to the T13 membership. It is reproduced below with Mr. Milligan's permission.

Some backgroud... Wren was a very popular family of disk drives made by Imprimis. The well know OEM customer was Compaq Computer. ST506 (ST412) was the interface used by Seagate hard disk drives (MFM and RLL drives). Seagate has a long history of using the word disc instead of disk.

The History of CAM ATA
by Gene Milligan

The original IDE was a proprietary implementation on a Wren II HH drive for a well known OEM customer. It was based upon emulating a controller board for ST506 drives, within the Wren drive. In part because it was a "customer special" not originally available to the general market it did not gain wide acceptance in the 5.25 inch arena. However shortly after its introduction 3.5 inch drives came along and the interface became the dominant 3.5 inch interface. Since it started as a proprietary implementation it did not have any specific name other than the product it was part of. The IDE terminology evolved later. (The acronym was not adopted as standards were written because it seemed confusing to the people working on the standard at that time. As we understood it the acronym either stood for IBM Disc Electronics or for Integrated Disc Electronics. Using IBM as a name was not attractive to some companies and not appropriate for a standard. Even in those days all disc drives we knew of, regardless of interface, had integrated disc electronics. So depending upon what you thought the acronym stood for it was either objectionable or not distinctive enough.

The small group working on the standard wanted to name it according to its then usage. As an unofficial industry group we chose to call it the PC AT interface. After our initial specification was agreed to and about to be proposed to the "ANSI" committee we realized that the PC AT terminology probably infringed upon trademarks and we needed a generic name. So we went for generic identity if there is any such thing as that. Consequently we changed PC AT interface to AT Attachment with the acronym ATA. Perhaps we were as guilty as the IDE folks since the acronym is more attractive than what it stands for. (Caution: In this Email to you I am not making any attempt to mark copyrights and trademarks. If you use any excerpts please add the appropriate credits.)

To back up slightly, while the unnamed interface took off in 3.5 inch drives there was still no reference specification. Consequently there were some compatibility issues when someone mixed and matched devices from different manufacturers. Probably the most frequent problem in the early days was failure to recognize some "slave" drives. At that time the "ANSI" peripheral interface committees were narrowly focused on a specific interface family under X3T9. X3T9.2 was putting together SCSI, X3T9.3 was putting together IPI, and X3T9.5 was putting together FDDI (X3T9.4 had disappeared when X3 rejected the IBM Channel). So I looked for a likely home to standardize the unnamed interface.

We had successfully standardized the ESDI interface in an unofficial industry group facilitated by Dal Allan of ENDL. This time I noticed that Dal was facilitating a group of system manufacturers, host bus adapter companies, and software companies to arrive at host standards for running SCSI devices. They were attempting to answer a number of the problems you describe in your Email. Some of the drive manufactures were also involved to understand and to help steer in a positive direction. They soon concluded that the host side could not be standardized on a hardware basis but probably could have hardware interoperability if they standardized on a Common Access Method (CAM) (I think of it as device drivers). So the committee became known as the CAM committee.

I decided that CAM might be a vessel of opportunity and requested in one of their meetings that they also address some of the compatibility issues in the unnamed interface. The bulk of the group would not have been very enthusiastic about this but a few of us quickly indicated that there was a small subgroup that were eager and willing to actually do the work if they would publish the specification under the CAM umbrella. That was agreed to and the I/O portion of the Wren II HH specification was donated electronically as a starting point. Most of the manufactures contributed to the completion of the original standard. That is why it has often, unfortunately being referred to as the CAM interface.

The ATA specification, as a competed spec was proposed to X3T9.2 to get the "ANSI" blessing. It was accepted and then cleaned up from a standards publication standpoint within X3T9.2. But implementations do not stand still. So the time for "standards cleanup" also required inclusion of various enhancements for new systems. In those earlier days we did not know how to put the period down. (Perhaps I suffer from that malady in this note.) Consequently the completion kept dragging out. Finally ATA was published as ANSI X3.221-1994.

Thanks to Gene for allowing us to use this!

Page updated 18 Dec 2005.