Rob van der Heij <[EMAIL PROTECTED]> wrote: >I am concerned about the approach of teaching adults to "download and >just run it to see what happens" That's one of the aspects that >distinguishes us from the folks who close their eyes and do "Next, >next, finish" and end up re-install their operating system on a >regular basis after a number of those actions.
>My preference is to use a familiar tool (like VMARC) to review what's >in the archive before I unpack it. And when I forget, I can at least >use it to get a list of the files it installed and use that to manage >the package. Transfer of a file in binary is not harder than transfer >in text mode - you just have to remember doing it right in both cases. >Accidental damage during transfer of a self-expanding program is most >likely to make it fail in a visible manner. Deliberate modification of >the code would be a concern. Somehow it does not make sense to me to >ship the packaging tools with each package. A self-extracting MAILABLE does NOT extract just by invoking itself -- when you invoke it, it tells you what's in it, same as a VMARC LIST. You have to say EXTRACT to extract it. I don't see how this is fundamentally different for the paranoid: "Here's a VMARC of a nice EXEC I wrote called CHRISTMA, that shows you a pretty picture of a tree" (or "Here's a VMARC of a new version of VMARC, no, really, it isn't malware, honest, would I lie to you?"). The "problem" with VMARC is that folks have to get VMARC itself. So that's an extra step, and one that can apparently be daunting for those with no experience transferring files across platforms (proof: if folks have trouble getting a flat, plaintext MAILABLE, you can be they'll have trouble getting a specific-format, binary VMARC). Transferring in binary *in a specific format* (F 80 in the case of a VMARC, F 1024 in the case of a COPYFILE (PACKed copy of VMARC MODULE) IS harder than plaintext, because different clients handle blocking/LRECL differently. This is the raison d'etre of MAILABLE. The truly paranoid can get a "known good" copy of MAILABLE and use *that* to extract MAILABLEs: MAILABLE some mailable a (LIST MAILABLE some mailable a (EXTRACT will both work on a self-extracting MAILABLE (it's smart enough to skip the Rexx code that's been prepended). So one conclusion from all this is that the biggest problem with MAILABLE is a lack of documentation discussing these issues. Which I'll fix soon, of course. And I'll include a discussion of LF vs CRLF, since that's been a problem for some people. Other suggestions welcome, of course... ...phsiii
