KROON@HGRRUG51.BITNET.UUCP (08/24/86)
Is there anyone out there who has experience with PSIAUTHORIZE? Does it work as you expected? Motive for these question is the following: When we first installed PSI (V3.2) I played somewhat with PSIAUTHORIZE but couldn't really get it to do what I wanted. I don't remember exactly what went wrong, but I remember one thing very cleary: I could let PSIAUTHORIZE crash with access violations, stack dumps and all these nice things, by feeding it a few relatively simple commands taken almost exactly from the samples in the documentation. I decided to forget about PSIAUTHORIZE because I believed we could do without it. I've learned better now from experiences that, as I believe, need no further explanation. I want to use the PSIAUTHORIZE facilities. I you found that it is unreliable I can stop trying and wait for a next release. If you are using it succesfully, could you please send me a copy of your PSI_SECURITY.COM (with scrambled DTE numbers and usernames of course) to find out what I did wrong? Thanks a lot. Peter A. Kroon.
carl@CITHEX.CALTECH.EDU (Carl J Lydick) (09/12/86)
You have to either specify what you want the product to do more clearly, or else let people know what the problem you're dealing with is. Having used PSI for a while now, my guess is that you've got everybody and his brother calling your site from all over the place, and reversing the charges, and that what you want to do is use PSIAUTHORIZE to allow you to permit reverse-charging from a restricted set of locations. If that's the case, then to the best of my knowledge, at least through version 3.2 PSI, you can't get there from here. Last November at the Anaheim DECUS symposium, one of the fellows who work for whatever part of DEC handles PSI was at a BOF, and one of the first questions that came up was, "is there any way to get PSIAUTHORIZE to do anything useful?" The guy spent about 10 minutes giving this vague question an equally vague answer; he was then asked specifically whether there was any way to allow reverse-charging from some but not all remote DTE's. It took him about 15 minutes before he decided we'd probably forgotten what the question he'd just failed to answer was, and invited the next. Version 3.2 PSI was just being released at about that time, and it turned out not to fix the problem, either. Now for the good news: Version 4.0 PSI is due out REAL SOON NOW (March, 1986, to be more exact; and yes, that is the right scheduled release date), has been released in Europe (as of August 13), had never been heard of by the local DEC office until I asked the software contracts manager there to get in touch with his opposite number in Geneva and verify that our site does indeed have a license for PSI, are in fairly desperate need of the new version (classes start here in about two weeks; version 4.3 VMS has at least 3 major bugs that I'm not at all happy with that get fixed in 4.4; and you can't run PSI 3.? with VMS 4.4; and I'd really like to go to 4.4 VMS before the 4.5 kit shows up), and have him authorize the manager of one of CERN's VAXes to send me a copy of the distribution. The snag there was that the folks in DEC's Culver City office never heard of EASYnet, either (which, considering that's DEC's own DECnet, and allegedly the biggest DECnet in the world, helps to explain why they never heard of PSI), and refused to believe that the DEC people in CERN would be willing to do any such thing. A week after my initial inquiry, they told me that the product was scheduled for release eight days after that, so since they ship by UPS, we should get our copy within ten days. That was about two and a half weeks ago. Sorry about the rather long flame, but DEC did essentially the same thing with PSI 3.0 about a year ago (due out in the spring; released in mid-summer; we managed to talk another site into sending us a copy in late summer, and got the official distribution in October). All in all, I don't have much hope that PSI 4.0's going to be much of an improvement, given the way DEC's handling it.