SE majors learn the how and why of release management, that's why it's relevant. IS majors learn business processes, and those tend to include release management and the importance of rollbacks.

It goes deeper than this.

Businesses have to pass down the costs of software maintenance to consumers; consumers won't pay what it actually costs to do this at the device level.

For servers, that's why you pay extra for long term support at a particular stable revision. See Red Hat Enterprise Linux and Ubuntu LTS, or Microsoft's long term support for older versions of Windows Server, for reference.

Developers of course would prefer to work on 'sexy' things like new features; maintaining older versions and backporting security updates and bugfixes is decidedly 'unsexy' in comparison.

Too many developers with CS degrees, too few with SE degrees, and nearly none with IS degrees. See http://www.acm.org/education/c... for more details on the differences.

According to the ACM (you know, the experts in this topic), there are five basic courses of study:
Computer Engineering (making the hardware)
Computer Science (designing the algorithms)
Software Engineering (release processes, patch management, etc)
Information Systems (translating business processes to code)
Information Technology (putting all the pieces of the system together an maintaining the whole lot).

You can read more at the ACM (Association for Computing Machinery) at http://www.acm.org/education/c...

Make sure you refresh those PROM's if they're EPROM or EEPROM (the absence of a window is no indication that it's a real fusible link PROM; it could be OTP UV EPROM in there). There is a thing called bit rot that occurs with most EPROM/EEPROM/Flash technologies where the isolated gate's charge bleeds off over time; 20 years is fairly normal, but 30 and 40 year old EPROMs (1702, 2708, and 2716 era) are beginning to fail all over. Search through the http://www.vintage-computer.co... forums as well as read the Wikipedia article ( https://en.wikipedia.org/wiki/... ) to learn more.

Mask ROMs are better, but not perfect. If the package is a 40 pin DIP it's almost sure to be flash, and that will bit-rot over time.

One more item on the checklist are those old paper caps that need to be replaced by X-class film on the inputs to power supplies. Again, the Vintage Computer forum is a great resource for information on how these things fail.

Also any batteries for NVRAM, like the ubiquitous Dallas Semiconductor devices, many of which are soldered in place. Or soldered in Lithium primary cells. Or like many older PC motherboards that have NiCd or NiMH cells that are both soldered in and leaking electrolyte. We have some Proteon routers (cisco's competitor back in the day) that have their NVRAM as low-power CMOS statis ram with a large bank of NiMH cells on a multibus card; they've long since lost any ability to retain charge.

As electronics age, lots of issues arise, and anyone who maintains such a system needs to see how others are handling the failures in these sorts of systems; again, the Vintage Computer forum is a great resource of talented people who are dealing with equipment of the same age. I know of many systems, particularly scientific instruments, where the controls are things such as a VAXstation 4000/90 connected to a SCSI CAMAC crate with wirewrapped boards, and VME Sun 2 and 3 series workstations controlling the whole lot. Keeping an aging VAXstation with VMS 5.2 or similar vintage running, with those old DEC StorageWorks 2GB and 4GB narrow SCSI drives, is a bit of a challenge, but when you have custom controls for multimillion-dollar equipment with no spares budget or major research instrumentation upgrade grant you have to get creative. (No, you can't just throw a PC in there, since the entire system's calibration depends upon the whole system timing and not just the actual platform). This system is being upgraded (there was even slashdot story about the upgrade at http://science.slashdot.org/st... ) but it's expensive to do things to the precision required.

Also, if the system uses GAL's or EEPROM-based FPGAs/CPLDs this is also something to make sure you have backups of the logic (JEDEC files, typically). Even fusible link PALs can go south. And be sure to have a stock of replacement chips, since many if not most of those older devices are long out of production.

Lots of test equipment is in this same boat, with expensive instruments like spectrum analyzers and the like running embedded MS-DOS and Windows on hard drives that are going on 20 years old. And, yes, in many cases they are consumer hard drives (I just looked at a very expensive 'multipath fading simulator' device, and it has a 6.4GB Western Digital Caviar drive in it.... you remember those? And one instrument I haven't looked into in a long time uses a 170MB Micropolis 5.25 full height ESDI drive.....

Don't know how many astrological discoveries there may be, since astrology looks into the future and this is looking into the past. But I do get the reference......

The difficulty with merging images could be the sparse nature of the data on the plate. In doing something like the visual6502.org dieshots there's plenty of data with which to do tiling; astronomical imagery is pretty sparse. So, while tiling will likely have to be done, the basic accuracy and precision of the platform (including the flatness and purity of the camera optics) is very important, and quite expensive.

Hebrew for 'Mountain.'

Yes. Required for the stability to scan with the precision and accuracy needed for both astrometry and spectroscopy. You need zero backlash positioners and a rock-solid (pun intended) surface.

Less expensive than the alternatives, such as refitting a PDS 2020G such as was used to generate Space Telescope Science Institute's digitized sky survey ( https://en.wikipedia.org/wiki/... ).

Yeah, reading Slashdot just isn't what it used to be. It, freshmeat.net, and linuxtoday.com were three of the sites I began reading in 1997 prior to putting up a new server install online..... of Red Hat Linux 4.1.

I forget when I registered, but it was a while after I had started reading it.

The only deal with a low id is that you have to maintain ancient e-mail addresses to keep it.....

So, the beta leaves a lot to be desired. But that's partly why it's beta, no? (and, no, I'm still on classic, and plan to stay that way as long as possible.)

The same thing that stops them from using Red Hat's src.rpms.... that is, 'nothing whatsoever.'

One of the big slowdowns for getting 6.0 out the door was getting 5.6 and 4.9 out the door.

I have rebuilt CentOS 5 from source on Itanium (IA64) and the step from 5.5 to 5.6 was rather interesting, and took a lot of thought as to which versions of certain libraries would build properly and needed to build properly in a very specific order (I don't recall right off hand the details, since it's been over a year since I did it, but there was a fairly substantial library uprev about halfway through the rebuild that had to be built after a few packages but before a few others; it was substantial enough that some of the packages in the first half would not rebuild at all with the newer version, and packages in the last half had to build with the newer version).

And EL6.0 was a beast to bootstrap, requiring a frankendist mix in the buildroots. I have not, and am not planning to, bootstrap it on IA64; Red Hat does do an IA64 for RHEL 5, but not for 6, so the source hooks and patches for IA64 were already in the source RPMS for 5, but are not for 6.

The members of the CentOS team have learned a great deal since then; I honestly think they got caught flatfooted by 6.0, but that's just my (very possibly incorrect) opinion.

Binary compatibility does not mean that the checksums match.

It means that every binary RPM has the exact same library version linkages and dependencies.

Yes, it should be good. I look forward to watching it, to see if either come up with any new arguments.

As to all the typical slashdot anti-creationism drivel, well, there are many logical and rational ways to talk about the various forms of creationism.

All science and math have to start out with postulates or axioms; things such as Euclidian geometry and spherical geometry, for instance, start out with axioms that simply cannot be proven, and these two exemplars each derive a useful system of geometry with contradictory axioms.

Evolution in general and neo-Darwinism in particular take as axiomatic that there is no creator and all things happened through the action of random chance, that is, the creation of order through stochastic processes, beginning with a single lifeform (common ancestry).

Creationism in general takes it as axiomatic that there was a creator, and things happened through that creator's initial action plus the actions of the various laws established through said creation.

In fact, it could easily be said that the creator was the Cosmic Egg (that of the Big Bang, you know, not that of various ancient middle eastern religions), since there is no way to go back past this event for which there is ample evidence (the cosmic background radiation, for instance).

It is impossible to disprove that a creator acted 6,000 years ago and made an old earth. It is also impossible to prove that a creator acted 6,000 years ago and made an old earth, too. For that matter, it is impossible to prove without any axioms that yesterday even existed. All proofs start with postulates, and all postulates and axioms are irrational. Irrationality is not a bad thing; just ask pi, e, and various square roots.

Axioms and postulates require faith in them, since they (by definition) cannot be proven.

As a thought experiment, put yourself in the postulated creator's shoes. You are getting ready to make the first trees; ok, how many rings to you put in them? Or in making a horse, what about the horse's teeth? Does the first man have a navel? All of those things are evidences of a past and of the passage of time; yet, if you create a tree today that has fifteen rings, your created man (created a couple of days later) could core into this tree and falsely state that it is 15 growing seasons old. You, acting as the creator, are making an old tree. Extrapolate to an old earth and an old universe.

My problems with evolution are that, even with the reams of evidence for microevolution, there are many more holes in the theories of macroevolution than there is evidence. For instance, the supposed primordial soup of the young earth can create amino acids; this much has been demonstrated in the laboratory. Oh, good, you have the building blocks of protein. Ok, mRNA can be synthesized in such an enviroment. That's good, now you have the blueprints for protein. But protein synthesis in even the simplest living cell requires more than a soup of assorted amino acids and mRNA (along with tRNA, rRNA, and DNA); other already synthesized protein 'machines' (ribosomes, for one) are required to do the kind of protein synthesis found in real cells (after all, a virus is basically those two things, a protein container with mRNA or other genetic material as a payload that hijacks already existing cells' protein synthesis machinery to build more viruses). Where is the evidence of each and every step required to make the first single celled organism? I would say 'simple' single celled organism, but in reality even the simplest single celled organism is massively complex.

There are lots of holes; until the holes are filled the theory is not proven.

Now, again, microevolution is readily observable and without doubt. But what about the most macro of the macroevolution foundation stones, the initial evolution of the first single cell? And what about that cell's reproduction? Mitosis is insanely complex at the molecular level.

So the evolutionist must postulate that the single cell 'just happened' (or one can cop out and say that aliens deposited the first life here, but then you're in creationist territory; not biblical YEC territory, but still solidly creationist). The spontaneous evolution of non-living material into a living cell has never been observed, nor is there a single fossil of such an event in the admittedly very sparse fossil record. I say spontaneous; if some scientist manipulates things in a laboratory to produce a living cell out of nonliving matter then that scientist has in fact become a creator..... and I say 'very sparse' about the fossil record because, relative to the geological timescale, there are precious few fossils, and the fossils that do exist show quantum leaps rather than gradual change.

Be open minded; what foundation does neo-Darwinism really rest upon? What are its axioms, its postulates that cannot be proven? Be honest about its holes, as there are many. And be open to conflicting points of view without dismissing them out of hand; this is one area where the open source movement is both a best example and a worst example: diversity is both encouraged and rabidly fought against in a great dichotomy. Agree to disagree if you must; but do it respectfully.

wsusofflineupdate ( www.wsusoffline.net )works well; use one machine to download everything, generate an ISO or USB key with the updates, and update offline.