lumpkin@nas.nasa.gov (Forrest E. Lumpkin) (02/26/91)
Below is some background of our workgroup and how we hope some day to use A/UX. If you don't care about this background and are only interested in my A/UX questions skip everything between the dashed lines :-) ------------ I've been dabbling with A/UX 2.0 here at work to determine if we should upgrade all of our networked macs from MAC OS to A/UX. Right now it seems that A/UX 2.0 is about 80% to 90% of what we want before we go ahead and spend the time and effort converting. To explain this assessment let me give some background of our work group. We are a group of scientists at NASA Ames doing research in hypersonic CFD (computational fluid dynamics). For the most part our work entails developing and employing computer models to study the physical models in which we are interested. We therefore use large amounts of computer time on CRAY supercomputers and do our development in FORTRAN. Before I get flamed for admitting that I'm a FORTRAN user keep in mind that we are scientists first and computer users second. (Also the best optimizing compiler for the Cray's vector architecture is the FORTRAN compiler.) On each scientist's desk is a MacII which is used for file editing, some simple graphics work, document preparation, as well as other general productivity tasks. We also have some SGI high end graphics workstations as well as a Sun 490 file server. Our interest in A/UX stems from the fact that all the machines in our network other than the Macs use some version of Unix. I therefore feel that A/UX has the possibility of making our work environment more heterogeneous (i.e. easier networking?, NFS support, etc.) as well as providing at the local level all the power of Unix for such things as managing files ( the Mac OS is very poor for managing large numbers of files - I have found Unix with its filters, pipes, shell scripts, regular expression pattern matching, and shell metacharacter pattern matching to be ideal for managing large numbers of files). The reasons why we have not abandoned the Mac OS are: 1) budgetary - we have mostly MacII's with a few IIcx's, so we well need hardware upgrades to run A/UX. As the funds become available we are slowly either replacing the MacII's or upgrading them so in time this constraint will disappear. (There is also the matter of buying an A/UX site license for 20 or so Macs - hopefully that is not going to be 20 times the $700 or so we payed one copy :-) ) 2) since we are scientists, not everyone in our group has the technical knowledge (or for that matter the desire) to become Unix system administrators. On the other hand managing the Mac OS is much easier in comparison and most of us are up to that task. The installation program that comes with A/UX 2.0 is the type of thing that will change this situation (although that program seems to have a few bugs as well as the bugs in HD Setup which hopefully will be worked out in future releases). When more user friendly tools for A/UX system admin. become available then I feel that most of us will not feel intimidated by the UNIX system administration task. 3) Everyone here likes the ease of use of the Mac OS. It is very likely that when/if we switch to A/UX many users will remain in a primarily Mac OS mode (i.e. using the Finder under A/UX) and resorting to Command Shell only when the ``extra power'' of the shell is required. -------- Given the above background I have a few questions/comments to pose to the A/UX community. 1) Will future releases of A/UX (2.0.1??) resolve the problem( undocumented feature:-) ) of only one Mac partition per physical device (hard drive). 2) Are user freindly A/UX sys. admin. tools in the works? Would other A/UX users like to see such tools/applications? (What I am suggesting is ``Maclike'' programs to do things like add/delete users, configure printing setups, create and mount file systems, etc.) 3) Will future versions of the Mac OS and A/UX merge even further? Two or three years down the road will one be able to ``ls'' his/her MacOS files from Command Shell (I know that this will require a MAJOR rewrite to the Mac OS file system, maybe something that could occur in System 8.0. Such a file system change in the Mac world is not unprecedented - there was the MFS to HFS evolution)? Is such a strategy something that the A/UX community would like to see? Is such a capability so technically unfeasible that it is never likely to occur? Let me conclude by saying I think Apple has a great product in A/UX 2.0; although I don't think it is quite mature enough for our whole workgroup. I'm going to continue evaluating A/UX on our one evaluation machine and I look forward to new advances on the A/UX front. Dr. Forrest E. Lumpkin III Computational Aerothermodyamics NASA Ames Research Center Mountain View, CA email: lumpkin@corvus.arc.nasa.gov