INFO-VAX Mon, 18 Jun 2007 Volume 2007 : Issue 329 Contents: Re: DECwindows Newbie Re: DECwindows Newbie Re: How will VMS be killed ? Re: How will VMS be killed ? Re: How will VMS be killed ? Re: How will VMS be killed ? issue with cpu usage on lock and unlock of system Re: Old UCX 4.0 manuals on line anywhere? Re: OpenVMS hobbyist license woes Re: OpenVMS hobbyist license woes Re: OpenVMS hobbyist license woes Re: OpenVMS hobbyist license woes Re: OpenVMS hobbyist license woes Re: OpenVMS hobbyist license woes Re: PLUG: PMAS Re: PLUG: PMAS Re: PLUG: PMAS Re: PLUG: PMAS Re: PLUG: PMAS Re: PLUG: PMAS Re: PLUG: PMAS Re: Question for the Group ---------------------------------------------------------------------- Date: Sun, 17 Jun 2007 18:10:22 GMT From: "John Wallace" Subject: Re: DECwindows Newbie Message-ID: "Robert Jarratt" wrote in message news:fRZci.96$KE1.18@newsfe1-win.ntli.net... > Second, if I click an application such as Clock on the Applications menu, it > takes me back to a login prompt. Why might this be? > Not sure if it would account for this particular symptom but you probably want to make sure that the login command files (system and user, and anything they invoke) are readable from your account with your default privileges, and that they aren't silently exiting prematurely e.g. due to something which doesn't work in a not-quite-interactive setup (the classic example would be "$ set terminal /inquire"). [Or am I confusing this with the classic cause of "network partner exited" messages from DECnet)?] hth John ------------------------------ Date: Sun, 17 Jun 2007 20:19:22 GMT From: "Robert Jarratt" Subject: Re: DECwindows Newbie Message-ID: "John Wallace" wrote in message news:iOedi.15569$aS5.6381@fe1.news.blueyonder.co.uk... > > "Robert Jarratt" wrote in message > news:fRZci.96$KE1.18@newsfe1-win.ntli.net... > > > >> Second, if I click an application such as Clock on the Applications menu, > it >> takes me back to a login prompt. Why might this be? >> > > > Not sure if it would account for this particular symptom but you probably > want to make sure that the login command files (system and user, and > anything they invoke) are readable from your account with your default > privileges, and that they aren't silently exiting prematurely e.g. due to > something which doesn't work in a not-quite-interactive setup (the classic > example would be "$ set terminal /inquire"). [Or am I confusing this with > the classic cause of "network partner exited" messages from DECnet)?] > > hth > John > > I am currently still logging in as SYSTEM and I have not yet customised any login command files. I see in SYLOGIN.COM that it seems to check the device type before doing a SET TER/INQ so it should not be that. In fact I commented out the line in SYLOGIN.COM just in case and I still have the problem. I still need to investigate some of the other suggestions I have received. Regards Rob ------------------------------ Date: Sun, 17 Jun 2007 14:10:00 -0500 From: David J Dachtera Subject: Re: How will VMS be killed ? Message-ID: <46758708.D7C43339@spam.comcast.net> Richard Maher wrote: > > My prediction is that if VMS has to listen to much more of the crap in COV > then it's simply gonna walk of the nearest tall building all by itself. ...and thank you for your support! -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Sun, 17 Jun 2007 16:15:36 -0400 From: JF Mezei Subject: Re: How will VMS be killed ? Message-ID: <4649c$46759675$cef8887a$19982@TEKSAVVY.COM> Marc Van Dyck wrote: > Didn't H.P. (or Compaq, I don't remenber) sign an agreement with the > U.S. D.O.D. stipulating that they would maintain OpenVMS alive for > 15 years or so ? No, those agreements were just garantees for support for 15 years. Consider that even with such deals, Compaq was able to savagely murder Alpha without warning. It can savagely murder VMS without warning too. The us military will probably port what it can port on its own, and for embedded systems in aircraft etc, will probably just require support and spare parts for the remainder of the DII/COE contract. Those embedded systems don't need VMS upgrades. ------------------------------ Date: Sun, 17 Jun 2007 15:29:30 -0500 From: Ron Johnson Subject: Re: How will VMS be killed ? Message-ID: On 06/17/07 15:15, JF Mezei wrote: > Marc Van Dyck wrote: >> Didn't H.P. (or Compaq, I don't remenber) sign an agreement with the >> U.S. D.O.D. stipulating that they would maintain OpenVMS alive for >> 15 years or so ? > > > No, those agreements were just garantees for support for 15 years. > Consider that even with such deals, Compaq was able to savagely murder > Alpha without warning. It can savagely murder VMS without warning too. > > The us military will probably port what it can port on its own, and for > embedded systems in aircraft etc, will probably just require support and > spare parts for the remainder of the DII/COE contract. Those embedded > systems don't need VMS upgrades. And with Charon-{VAX,AXP}, they might not even need spare parts... -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Mon, 18 Jun 2007 07:17:35 +0800 From: "Richard Maher" Subject: Re: How will VMS be killed ? Message-ID: C'mon David, > ...and thank you for your support! I think you'll find that my (very occasional) motivational pep-talks with VMS engineering contain insightful and poignant commentary; it what everybody else writes that is "crap" :-) Cheers Richard Maher "David J Dachtera" wrote in message news:46758708.D7C43339@spam.comcast.net... > Richard Maher wrote: > > > > My prediction is that if VMS has to listen to much more of the crap in COV > > then it's simply gonna walk of the nearest tall building all by itself. > > ...and thank you for your support! > > -- > David J Dachtera > dba DJE Systems > http://www.djesys.com/ > > Unofficial OpenVMS Marketing Home Page > http://www.djesys.com/vms/market/ > > Unofficial Affordable OpenVMS Home Page: > http://www.djesys.com/vms/soho/ > > Unofficial OpenVMS-IA32 Home Page: > http://www.djesys.com/vms/ia32/ > > Unofficial OpenVMS Hobbyist Support Page: > http://www.djesys.com/vms/support/ ------------------------------ Date: Sun, 17 Jun 2007 22:03:44 -0700 From: sowmya.rangineni@gmail.com Subject: issue with cpu usage on lock and unlock of system Message-ID: <1182143024.615752.100020@z28g2000prd.googlegroups.com> Ours is a windows based application. When we open the application the CPU usage is 0% and the Memory Usage is 54,324Kb When I open a specific form in a module, the CPU usage is 0% and the Memory Usage increases accordingly (67,730 Kb). Now when I lock and unlock the System the CPU usage increases to 50% with the Memory Usage being 67,730 kb. I have made use of a .Net memory profiler and had taken snapshots before locking and unlocking the system. The following is the result after comparing both the versions. NameSpace Name Total New Removed Delta Total Max Min Delta System WeakRefrence 1681 613 0 613 26896 16 16 9808 System.Reflection RunTimeMethodInfo 1229 8 0 8 29496 24 24 192 System.Windows.Forms NativeMethods.TracKMOU.. 10 5 0 5 240 24 24 120 System Runtimetype 2725 4 0 4 4360 16 16 64 System String 37232 5 1 4 2256 102456 18 458 Can anybody let me know the reason for the increase in CPU usage and how I can overcome this. And why is there a drastic increase in the number of instances for WeakReference class when the operation is just locking and unlocking and nothing to do with the application. Thanks in advance Sowmya ------------------------------ Date: Mon, 18 Jun 2007 00:42:03 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: Old UCX 4.0 manuals on line anywhere? Message-ID: In article , burley.not-this@encompasserve-or-this.org (Graham Burley) writes: > > >In article <1182089044.395469.71590@c77g2000hse.googlegroups.com>, Charlie writes: >> >>> The crux of the documentation is in BookReader format. >> >> any chance that someone can point me to bnu for windows? What I've >> read is that I need to order a doc cd to get it, which seems a bit >> silly. >> > >I copied those to Deathrow, you can view courtesy of HyperReader at: > > http://www.openvms-rocks.com/docs/network/ucx41/dy9mzaa3.odl > >Looks like 4.1 docs, but the majority (e.g. Management) are 4.0. Yeah, go figure. They all came off of the same quarterly ConDist. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" ------------------------------ Date: Sun, 17 Jun 2007 20:07:57 +0200 From: Wilm Boerhout Subject: Re: OpenVMS hobbyist license woes Message-ID: <4675787d$0$25490$ba620dc5@text.nova.planet.nl> on 17-6-2007 19:54 rtk wrote... > 1. TCP/IP doesn't run automatically on boot, I have to run TCPIP > $CONFIG.COM and manually start it. I think I should be able to add > something to the STARTUP.COM file to run it but I'm not sure what to > add. Do not touch SYS$SYSTEM:STARTUP.COM, instead, Add the line @SYS$STARTUP:TCPIP$STARTUP to SYS$MANAGER:SYSTARTUP_VMS.COM > 2. I added a new user per instructions above, got no error messages, > but it won't let me logon as that user. I seriously doubt that I need > to restart the machine or anything like that to get a new user > recognized. I haven't really looked into it much yet, just a comment. No restart should be necessary. What error message when trying to login as that user? > 3. What's the proper magic to get the CD-ROM mounted again. It is on > DKA400: and that's how I booted it to install in the first place. > Again, I haven't dug into it yet, it is Father's Day after all. Insert a VMS-formatted CD, then $ MOUNT /OVER=ID DKA400: Have fun! /Wilm ------------------------------ Date: Sun, 17 Jun 2007 14:08:06 -0500 From: David J Dachtera Subject: Re: OpenVMS hobbyist license woes Message-ID: <46758696.287309D2@spam.comcast.net> rtk wrote: > > On Jun 17, 8:38 am, David J Dachtera > wrote: > > Again, I'm thinking of using your experience to build an OpenVMS Hobbyist > > Beginner's FAQ to cover such questions. Perhaps over the course of the summer, > > if time permits. > > This would be wonderful. I think I'm almost there but have to > struggle with little things: > > 1. TCP/IP doesn't run automatically on boot, I have to run TCPIP > $CONFIG.COM and manually start it. I think I should be able to add > something to the STARTUP.COM NO!!! You add site-specific items to SYS$MANAGER:SYSTARTUP_VMS.COM, *ONLY*! > file to run it but I'm not sure what to > add. $ @sys$startup:tcpip$startup ! Start TCP/IP Services for OpenVMS Take a look at SYSTARTUP_VMS.TEMPLATE in the SYS$MANAGER path. > 2. I added a new user per instructions above, got no error messages, > but it won't let me logon as that user. I seriously doubt that I need > to restart the machine or anything like that to get a new user > recognized. I haven't really looked into it much yet, just a comment. When adding a new user, you're actually COPYing from the record called DEFAULT. If you do SHOW DEFAULT within AUTHORIZE, you may notice that in the "Flags:" line it probably shows DISUSER; that is, that specific record is disabled from any logins. The (inobviously documented) syntax to enable your new user record would be "SET username/FLAG=NODISUSER". > 3. What's the proper magic to get the CD-ROM mounted again. It is on > DKA400: and that's how I booted it to install in the first place. > Again, I haven't dug into it yet, it is Father's Day after all. It depends. Generally, however, this will work: $ MOUNT/OVERRIDE=IDENT DKA400: Now, that will MOUNT it privately to your process (login session, batch job, etc), not system-wide. When you're done using it, DISMOUNT it, or just logout. See, unlike smaller-scale operating systems like DOS or Windows, VMS and the o.s.-es from which it descends do not automatically assume that any "visible" device is for the system's exclsuive use (read up on VMS Clusters to get some idea of why this done) or is, in fact, ready to be used. Therefore, you must "give the system permission" to use a device by MOUNTing it, similar to the UN*X "mount" command. See HELP MOUNT for more info. VMS is a commercial operating system. Commercial o.s.-es are a somewhat different paradigm from what you may have encountered thus far. > Just some fodder for your FAQ. I'd assume your readers to be people > like me who might have been casual VMS users a decade or two ago and > now have come across old hardware that they are trying to resurrect > and who don't need to ask such silly questions about other operating > systems ;) Welcome to the VMS fold! Apologies for much of griping here in this newsgroup, by the way. > Another thing for your FAQ: a guide to essential manuals. I have > about 30-40 OpenVMS manuals and about 15-20 VAX VMS manuals in my > garage, for example, but it is hard to tell where to start (I am > looking at the OpenVMS User's Guide now). I had to take the manuals > if I wanted the machine :) The System Manager's Guide is likely in one small (5 inches or so by 8.5 inches) manual, and/or two larger manuals: one is labelled "Essentials", and the other has to do with larger and more complex configurations. You can find current versions on-line at: http://www.hp.com/go/openvms/doc ..., and then click the lick in the left-hand column that says, "OpenVMS Operating System". > I did configure my router so that telnet/ftp requests are sent to the > Alpha while ssh/http requests get sent to my Linux server (good guess, > that's my primary os). That's pretty neat as I can get to the Alpha > now from anywhere. Should be safe enough, I would think. To a point. Anything that responds to FTP and/or Telnet connections will receive great deal of "hammering" by "crackers". They typically try usernames/passwords that are found on Windows (Example: Administrator, Guest) and UN*X (Example: root, nobody). VMS, however, employs breakin detection and evasion to virtually eliminate the possibility of getting cracked. This does not stop the attacks, but ensures that a successful login is highly unlikely, even in the event that a password is successfully guessed. VMS has been banned from Hacker conventions as "uncrackable". > I recall using a CD.COM file long ago that acted like the DOS cd > command. I might have it somewhere, but if it were a common thing > might anyone here still have it? And, where would be the best place > to put these things so that they can be used from anywhere? Is there > a repository of cool/useful DCL files online anywhere? Look at http://www.djesys.com/freeware/vms/ and pull down the 4038_freeware.zip file. There's a CD.COM in there that I use, and other useful goodies. (Must be UNZIPped on VMS, however - neither Windows nor UN*X understand RMS file attributes and formats without add-ins.) > Now that FTP is working I can load all the licenses at once. I assume > there will be no trouble if I reload a license for something already > licensed or if I register licenses for things that are not installed > yet. Correct. I usually load licenses using a terminal program and the serial console port, and just cut/paste between my e-mail user agent and the terminal program. There was a license management thread a while back - some folks' hobbyist licenses expire in the early part of the year. Try to Google for it. Lots of good license tips there. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Sun, 17 Jun 2007 16:31:12 -0400 From: JF Mezei Subject: Re: OpenVMS hobbyist license woes Message-ID: Wilm Boerhout wrote: > > Insert a VMS-formatted CD, then $ MOUNT /OVER=ID DKA400: > This will mount the CD privately (your own process). And you need to add /NOWRITE to CD media. You can mount it systemwide with MOUNT/NOWRITE/SYSTEM dka400: volumelabel logical name for instance MOUNT/NOWRITE/SYSTEM dka400 VAXDOCJAN952 $disk3 you can then refer to the disk as $disk3 (eg: DIR $DISK3:[000000] ) If you have a cluster, just add /CLUSTER to the mount command and the disk automatically becomes mounted with that logical name on all nodes in the cluster. ------------------------------ Date: Sun, 17 Jun 2007 21:13:52 -0000 From: IanMiller Subject: Re: OpenVMS hobbyist license woes Message-ID: <1182114832.789422.254380@q75g2000hsh.googlegroups.com> you may find the pointers at http://www.openvms.org/pages.php?page=Beginner http://www.openvms.org/pages.php?page=Books http://www.openvms.org/pages.php?page=Documentation useful to get you started. ------------------------------ Date: Mon, 18 Jun 2007 01:11:01 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: OpenVMS hobbyist license woes Message-ID: JF Mezei writes: >Wilm Boerhout wrote: >> >> Insert a VMS-formatted CD, then $ MOUNT /OVER=ID DKA400: >> >This will mount the CD privately (your own process). And you need to add >/NOWRITE to CD media. VMS is smart enough to figure out that a CD is nonwritable without needing to be told with /NOWRITE. ------------------------------ Date: Sun, 17 Jun 2007 21:18:19 -0400 From: JF Mezei Subject: Re: OpenVMS hobbyist license woes Message-ID: <6499c$4675dd67$cef8887a$32523@TEKSAVVY.COM> Michael Moroney wrote: > VMS is smart enough to figure out that a CD is nonwritable without needing > to be told with /NOWRITE. For a DK device you are right. But with a different type of controller that makes a CD appear as a DUAx device, VMS doesn't know it is a CD and the mount will fail without /NOWRITE (sometime between 5.5.-2 and 7.2 on vax). ------------------------------ Date: Sun, 17 Jun 2007 18:24:14 +0000 (UTC) From: david20@alpha1.mdx.ac.uk Subject: Re: PLUG: PMAS Message-ID: In article , "John E. Malmberg" writes: >Bill Gunshannon wrote: >> In article , >> helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: >> >>>In article <5djfl4F3575eiU1@mid.individual.net>, bill@cs.uofs.edu (Bill >>>Gunshannon) writes: >>> >>> >>>>So, what is the technological solution? >>> >>>ZEN.SPAMHOUSE.ORG. It's an RBL. And it works fine with HP TCPIP (as of >>>version 5.4). >> >> >> And how much legitimate business have you lost because someone who >> wanted to contact you had their email bounced because of an RBL for >> reasons totally out of their control? It may stop some SPAM, but >> it also stops some legitimate traffic. And you will never know you >> lost that sale!! > >With the current state of art of spam filtering, the real state of the >art, not the one promoted by many commercial spam filtering, the error >rate of the most popular and free DNSbls is almost too low to be measured. > >It is far lower than the rate of human error, or from e-mail lost >through various network and server failures on the Internet. > >Current state of the art for using something like the zen list from >spamhaus has resulted in statistics from multiple sources of from about >80 to over 90% of spam delivery attempts blocked and no false positives. > By their very nature DNSBLs block all mail whether legitimate or spam coming from the blocked address. To some extent this is a good thing since it means the legitimate users put pressure on any ISP getting itself on a blacklist to clean up it's act. However it does mean that legitimate mail from those sources is blocked until the ISP has managed to get itself off the list. I think this is probably what Bill was referring to rather than implying that blacklists like zen.spamhaus.org publish incorrect information (though since the policies of different blacklists vary so much you need to be careful to make sure that whichever ones you implement are correct for your systems and users). >I have been monitoring forums where people complain about incorrect >blocking of DNSbls very loudly for quite a few years now. > >In the past 5 years, I have seen 0 (zero) complaints about an incorrect >listing on a spamhaus list. > >All the ones that I have seen about open relay and open proxy lists have >turned out that there really was a severe problem on the mail server >that the owner refused to admit or fix until they found almost no one >accepting their e-mail. > >Even on the aggressive spamcop.net list, in the majority of the cases >where someone complains about an incorrect listing, the investigation >shows that criminals and spammers had more control of the mail server. > >In the past 10 years, I have had only two people claim that their >attempt to e-mail me was blocked because they sent it from a mail server >operating on a DHCP address. And in both cases, their ISPs had publicly >readable TOS that stated that it was prohibited for those I.P. addresses >to be running any servers. > >Because other large ISPs are starting to totally block other ISPs that >allow zombies to send spam from their DHCP ranges, most residential ISPs >are planning to implement port 25 blocking for those ranges. And in >many cases such blocks are implemented with out notice to their >customers in order to get a block removed or to prevent a block from >being put in. > >> Realize, with my suggestion you can turn this whole thing around. >> Once you have all your regular business on a trusted network you >> can reverse the logic of your SPAM filtering. Start looking at >> email that comes in from the unttrusted network (ie. The INTERNET) >> with filters looking for keywords you want to see, like your product >> name or somthing else that doesn't look like "Viagra". :-) And, >> dump the rest. > >Keyword filtering even with Bayesian filtering has been shown to be >totally ineffective at spam filtering at a global basis. Every system >that I have seen implemented, has been totally ineffective at detecting >the majority of spam, including ones that are easily detected because of >header defects that are *NEVER* found in legitimate e-mail. > >The only content filtering that I have seen that is 100% accurate in >detecting the presence of spam in a message is the URL lookup, where it >checks to see if a URL resolves to an IP address in a conservative >blocking list. And you only want to do that check if there is a header >defect, the source is a public web-mailer, or the source is on an >aggressive list like a multihop or an aggressive list like spamcop.net. > >Combining the URL check on the small percentage of spam that gets >through a good selection of conservative blocking lists can get the spam >leakage down to far less than 1 percent. Again with less risk of >rejecting a real email than human error of just hit delete. > >The thing to remember is that anti-spam systems that are cause detected >spam to be rejected have visible "error" rates, and overwhelmingly those >cases are from severe problems with the sending system. > >Unfortunately what is far more visible than the success stories, mostly >with open source and free products, is products in the commercial sector >that do not work and are fundamentally flawed in their operation. > >Any system that does not issue SMTP rejects for detected spam is flawed, >because it will have silent errors. > >Basically where I see the effort right now in the anti-spam community is >a growing war against "backscatter" where mail servers are sending new >bounce messages for detected spam and viruses instead of SMTP rejects. >This war has been pushed because some of the low end and cheap >commercial mail servers can not do spam and virus filtering properly, >and at least one major "anti-spam/anti-virus" product comes out of the >box configured to send a backscatter message to any thing it detects. > >That an anti-spam product even has that as an option, let alone a >default, means that the author has no clue about how spam and viruses >exploit weaknesses in the SMTP protocol, and you do not want to have >such a thing touching your e-mail if you care about it. > >Many small networks have had their mail servers or other equipment >overloaded from such backscatter, and it is very common for dial up >users to find that they are losing legitimate e-mail because their mail >quota was full of backscatter from a virus or spam. > >The social solution to the spam problem is to expose the bad practices >in mail server operation for what they are and to refuse e-mail from >those networks that permit such things to happen. > >And this is happening more with large networks and private blacklists >than it is with public ones. But it is happening with both. > >I think that Hunter has posted here that the PMAS as a proxy server >rejects the detected spam before it gets into the mail server. As I do >not operate a mail server (It is prohibited by the class of service I >have from my ISP) I have not actually tried the product. > Pretty much all mail systems reject blacklisted entries early on in the smtp dialogue. However not all systems do when content scanning. PMAS in proxy server mode does do this - and I have asked Process to update PMDF so that in the future PMAS working directly with PMDF can do the same. However rejecting spam in that way only really works with the threshold set fairly high - ie when you can be almost certain that it is definitely spam. If you set the threshold too low then you get a fair amount of legitimate mail rejected and the legitimate sender then is left wondering why you are labelling him as a spammer and what he can do to change his mail to get it past your "defective" filters. Unfortunately passing back enough information from anti-spam products to let the sender really know why his mail was rejected so that he can correct it is very difficult (if not impossible). So at the moment the best solution seems to be 1) Reject mail using DNSBLs during SMTP dialogue 2) Reject mail with a high threshold spam score from an anti-spam product during the SMTP dialogue if possible. 3) Quarantine or tag and deliver more problematic spam (ie that scoring as spam but with a lower threshold score). David Webb Security team leader CCSS Middlesex University >-John >wb8tyw@qsl.network >Personal Opinion Only ------------------------------ Date: Sun, 17 Jun 2007 19:26:02 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: PLUG: PMAS Message-ID: In article <5dkqn3F350el9U5@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: > And how much legitimate business have you lost because someone who > wanted to contact you had their email bounced because of an RBL for > reasons totally out of their control? It may stop some SPAM, but > it also stops some legitimate traffic. And you will never know you > lost that sale!! If you look at the reasons sites get listed in the RBL, you will see that no-one doing legitimate business would send email from such a site. ------------------------------ Date: Sun, 17 Jun 2007 19:28:57 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: PLUG: PMAS Message-ID: In article , david20@alpha2.mdx.ac.uk writes: > However if things are setup correctly the sender should get back a message > saying that they have been blocked because they are on that particular RBL. > The idea is that the legitimate user can then complain to their ISP and > persuade them to take action to make sure they are not on such RBLs. > Since the rejection is done during the SMTP dialogue rather than after > accepting the mail message the bounce with this reason is much more likely to > get back to the legitimate sender. Indeed. In SMTP.CONFIG, you can customise the text. I point them to a URL with a longer explanation. > However using content scanners rather than RBLs provides it's own problems in > this regard. > The problem with deleting messages after they have been inspected by > anti-spam content scanners is that the mail is silently discarded (since the > scanning is generally done after the message has been accepted). Right. And one shouldn't bounce them, because the sender is probably faked. ------------------------------ Date: Sun, 17 Jun 2007 13:52:48 -0700 From: DeanW Subject: Re: PLUG: PMAS Message-ID: <3f119ada0706171352o747f7c10n4b67ab50262039b8@mail.gmail.com> On 6/16/07, Phillip Helbig---remove CLOTHES to reply > That seems about average. I've resorted to using ZEN.SPAMHAUS.ORG as an > RBL. That gets rid of the lion's share. I use ASSP (assp.sourceforge.net), which is an effective stand-alone proxy. The two features I appreciate the most: 1) Delaying- the first time it sees a message from IP, from USER, to RECIPIENT, it returns a "temporary failure" and logs the triplet. If that triplet comes up again in < 5 minutes, it gets rejected again and logged as a spammer. If more than 5 minutes, then it's considered a valid sender, and logged as such; future messages are not delayed (unless it fails one of the subsequent spam checks). If it doesn't come back in 24 hours, the entry is purged. 2) Max errors: To defeat dictionary attacks, after 3 invalid recipients, the connection is dropped. Those two block 87% of inbound spam before the message body even begins to be transmitted; RBLs and a Bayesian filter catch almost all the rest. Each user here sees 1-2 spam messages a day. I can live with it. ------------------------------ Date: 17 Jun 2007 17:24:39 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: PLUG: PMAS Message-ID: In article , david20@alpha1.mdx.ac.uk writes: > By their very nature DNSBLs block all mail whether legitimate or spam > coming from the blocked address. No, by their nature, DNSBLs only _list_ those IP addresses. What happens next depends on the software at the recipient end _using_ the DNSbl: 1. MUA software at best can segregate email from the listed IP address, making it available for possible (but unlikely) review by the recipient. 2. MTA software _can_ do the above, but it can also do the more reliable thing - REJECTing the email in the SMTP dialog, so the sender of legitimate mail knows it did not get through. Some broken software for MUA's tries to send a mail back to the originator saying the mail was not given to the recipients, but since most spammers falsify the "From:" header, the result is the broken software has generated new spam, sending a non-delivery report to an innocent bystander. ------------------------------ Date: Sun, 17 Jun 2007 22:28:17 +0000 (UTC) From: david20@alpha2.mdx.ac.uk Subject: Re: PLUG: PMAS Message-ID: In article , helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: >In article <5dkqn3F350el9U5@mid.individual.net>, bill@cs.uofs.edu (Bill >Gunshannon) writes: > >> And how much legitimate business have you lost because someone who >> wanted to contact you had their email bounced because of an RBL for >> reasons totally out of their control? It may stop some SPAM, but >> it also stops some legitimate traffic. And you will never know you >> lost that sale!! > >If you look at the reasons sites get listed in the RBL, you will see >that no-one doing legitimate business would send email from such a site. > I took Bill's usage of RBL to be a reference to any DNS blocklist rather than just MAPS RBL. RBL is often used that way - though it probably shouldn't be. The listing policies of different DNS blocklists vary greatly. David Webb Security team leader CCSS Middlesex University ------------------------------ Date: Sun, 17 Jun 2007 23:16:42 +0000 (UTC) From: david20@alpha2.mdx.ac.uk Subject: Re: PLUG: PMAS Message-ID: In article , Kilgallen@SpamCop.net (Larry Kilgallen) writes: >In article , david20@alpha1.mdx.ac.uk writes: > >> By their very nature DNSBLs block all mail whether legitimate or spam >> coming from the blocked address. > >No, by their nature, DNSBLs only _list_ those IP addresses. What happens >next depends on the software at the recipient end _using_ the DNSbl: > > 1. MUA software at best can segregate email from the > listed IP address, making it available for possible > (but unlikely) review by the recipient. > > 2. MTA software _can_ do the above, but it can also do > the more reliable thing - REJECTing the email in the > SMTP dialog, so the sender of legitimate mail knows > it did not get through. > >Some broken software for MUA's tries to send a mail back to the >originator saying the mail was not given to the recipients, but >since most spammers falsify the "From:" header, the result is >the broken software has generated new spam, sending a non-delivery >report to an innocent bystander. I was responding to " >> And how much legitimate business have you lost because someone who >> wanted to contact you had their email bounced because of an RBL for >> reasons totally out of their control? It may stop some SPAM, but >> it also stops some legitimate traffic. And you will never know you >> lost that sale!! " which talks about lost legitimate mail because of bounces because of an RBL. I'm fully aware of how DNSBLs work but in the context the form of my reply seemed appropriate. I'm sorry for using bad terminology. The point I was making is that the DNSBL just supplies the IP address and hence any action taken is just based upon the fact that the email came from that IP address. The MTA using the DNSBL cannot distinguish between legitimate mail coming from that IP address and Spam mail coming from that IP address without passing that mail message through some other check - such as a content scanner. Hence once the address is listed because of some misuse then all legitimate as well as spam mail from that address is affected until the address is delisted. To some extent this is a good thing since it means that the legitimate users put pressure on any ISP getting itself on a blacklist to clean up it's act. David Webb Security team leader CCSS Middlesex University ------------------------------ Date: Sun, 17 Jun 2007 20:12:53 -0700 From: AEF Subject: Re: Question for the Group Message-ID: <1182136373.566444.243110@g4g2000hsf.googlegroups.com> On Jun 13, 4:46 am, Michael Kraemer wrote: > AEF schrieb: [...] > > Again you're justifying letting the patient die by saying, "Look, the > > patient is dying. Let's go home." It's the POTENTIAL profits. High- > > margin profits. > > > History has proved that companies are willing to spend more for VMS if > > it weren't allowed to languish as it is. Would it kill HP to try? > > Would the shareholders revolt if profits from VMS increased? Doesn't > > it make sense to market your high-margin products? > > Well, I'm not in HPs shoes and I have no business > in defending their actions, I just find their strategy comprehensible. > Why should they put extra effort on such a tiny fraction of their business ? Because VMS systems have higher margins. Suppose that HP replaced all its customers' Windows systems with VMS systems. This would cause a big increase in profits, no? This would have the added advantage of not having to deal with MS or Linux. And HP would have a hold on the customers. And the customers would greatly benefit from a very superior operating system. What horrors am I missing in this scenario? And how much extra effort would it really take? Just take a reasonable fraction of VMS's profits to do some marketing and see what happens. Would it kill anyone to do so? > Even if VMS revenue was several times larger at the time of > the Compaq takeover, it still was dwarfed by HPs ink ocean. So were its calculator revenues. Should they let their calculator business whither away? So why don't they just drop everything else and re-brand themselves as INKS R US? They could rewrite their motto: The hell with invent. Ink it up! > As far as the potential goes: I think (as does probably HP) > you vastly overestimate that, the hey-days of VMS are long gone. Again you repeat the self-defeating logic: "Why fix something that's broken?" Well, you don't fix something that's not broken!!! So if VMS is so lame, why doesn't HP sell it? In fact, why don't they sell the entire BCS division if it's such tiny a portion of HP's revenues or profits? (I don't really know how big BCS is in HP, so maybe ignore this last statement. Correction welcome.) > The few companies who still believe VMS is indispensable for their > IT will buy it anyway, without marketing. That's EXACTLY why VMS needs marketing: To be sold to people who may not find it "indispensable", but to whom VMS would still be a better choice. ... Hey, VMS is always the better choice, no? :-) Yet again you return with your self-defeating "Why fix something that's broken?" logic. Back in 2000 I interviewed at Prudential. They said they were looking for a VMS sys admin but that their VMS systems were going to be decommissioned in a year or two as they switch to some Unix solution. They give one reason as Oracle insisting they run on one particular version of VMS. I think there were other reasons. Anyway, if DEC/ Compaq/HP hadn't killed the Alpha and had actively promoted VMS they'd still probably be using it. By resuming promotion and increasing support, HP can win new VMS customers like Prudential. And it would help sell more Itanium systems! > Its (almost) just like IBM treat their "legacy" iSeries/zSeries business. I'm not familiar with that. But I did work with the IBM AS/400 and its OS/400 operating system back in 1997. Switching among other operating systems is sort of like switching car makes and models. The controls are a little different, and the labeling may be different, but it's sort of the same thing and it's not too hard to get going at least at a beginner's level. But switching to the OS400 (OS/400?) operating system was more like being plopped into a parallel universe with different laws of physics where nothing works as it did before and there is nothing familiar to get your bearings from. On the bright side, the screen was very tightly controlled and very neat. Function keys were always defined on the screen as you moved among different screens and functions. No burst of anything randomly in the middle of the screen. No lost messages, IIRC. But it was really hard to notice the really tiny new-mail indicator. IBM seems to put a lot more into their operations control systems than other companies. This also reflects my experience using IBM computers at the New York Blood Center in 1994. AEF ------------------------------ End of INFO-VAX 2007.329 ************************