If someone were to 'extend' the way that an LDAP consumer interacts with a supplier, there's an issue. As it is now, every set of LDAP software has its own way to handle direct replication of data.
However, the type of information that can be placed into an LDAP directory is handled by the schemas.
In much the same way, a database server doesn't care what your table definitions are to maintain ODBC compliance.
Now, in this case, we have schemas, which much like a table definition, define what sort of information is allowed to be placed into the directory. It may be mail addributes, such as 'rfc822mailalias' and 'mailforwardingaddress', it might be posix authentication information, such as 'loginshell' and 'homedirectory'.
In this case, I have no idea what schema extentions OS X may require, but the standard posix attributes would make sense (ie, what you'd find in/etc/passwd normally).
Damn, and people thought I'm strange for making bets on when the movie actually starts. (I'm doing best w/ 12 minutes after the posted time for the movie to start, but I've noticed that sometimes the day the movies come out, there's only about 9-10 minutes of crap).
Oh...and before you make bets, make sure everyone agrees on the clock that you're using. I normally use my cell phone, as keeps better time than wrist watches.
Let's see, there's um....wait, no, that's command line.
There's um....wait, no, that's SunOne, which is really iPlanet, which was the Sun/Netscape Alliance, which was Netscape...
um.... well, the OS installer doesn't suck that bad...assuming you're working from a jumpstart server, and don't need to work interactively.
[Oh...from experience...you're still better off trying to install it yourself than letting Sun Professional Services do it for you... installed all the applications in/, and configured SIMS to use one large mail partition, without mentioning to us that the SIMS backup program stops when it hits a 2G limit, and that's the only reason we were making our 4hr backup window. [over 300G of data]]
ISPs are value-added resellers, who sell a product based on queuing theory.
What you seem to be confused about is the two parts to the charge, which normally isn't seperated out on DSL and other smaller connections.
There's line, and there's port.
Line is the charge for the connection between you, and the ISP. Sometimes there's markup, sometimes there isn't. There's normally markup if the ISP owns the equipment used to terminate the line. No real queuing here when you're dealing with a form of dedicated circut, which would be basically anything but standard dial-up.
Port is the charge for your data to reach the internet. You always have to deal with queing theory here, or well, everyone gets screwed. You may have a max of 1.5kb download, but is there any sort of a CIR on the line? [what's the speed that they'll guarantee you to get? I'm guessing, nothing, as it's not in your contract with them].
Take the example of the ISP that I used to work for [before they started doing DSL]. We had somewhere near 200 modems per T1. With all 200 people using a 56k modem at the same time, that'd be over 10mb...but a T1's only 1.5mb. In our case you could always assume a certain amount of slack on the lines -- even when a person's connected up, it's not working full tilt, or it's doing something local, like downloading e-mail.
Most ISPs have something to the effect of requiring 'interactive' sessions or something similar, to keep you from sucking down the bandwidth with servers or bots. If you wanted to run a server, upgrade your plan, and get a business plan, or dedicated dial-up, or something similar.
People who try to get stuff for free, and take the ISPs for a ride are the reason that the market crapped out. You have a contract of service with your ISP. If you are using services which are specifically denied by your AUP, you are stealing services from them. It's no different than going to a store and lifting something. Just because it CAN be done, doesn't mean that it should be done.
Oh...and all ISPs should just be banning all incoming traffic on known ports if they have any hopes of staying alive.
We had to have a forced power outage while they were installing power cutover, so they could service the UPSes in our rather aged datacenter.
As we had between 2-4 hours that we were going to be down, as we hadn't gotten approval for a maintenance window in over a year, we moved some disk arrays around, replaced out some old sun "Mass Storage" racks [30" wide cabinet, 28RU, single power] for "Expansion Cabinets" [24" wide, 36U, dual powered].
In the process, we also moved a few systems that were in a rack that was primarily filled with a system going out of production, and cleaned up a signficant chunk of the rats nest of cabling behind these cabinets.
We were thanked by the data center management for making everything look nicer. [Not placing the systems with redundant power supplies in a rack with redundant power supplies, preventing future downtime for systems, creating more room in the machine room, etc, etc, etc] Hell, we weren't even directly thanked for pulling it off with less than 2 weeks notice, and only losing one system in the process [Sparc Storage Array 200, which had been in service for over 6 years with no signficant downtime]
Oh...and even more important than looking nice -- it helps if it actually fits through the door, and doesn't cause the floor to buckle.
Personally, I work for a university, so I get to take classes, and they're rather flexible about my taking classes in the middle of the day, so long as it doesn't affect a production service.
First off, remember that your schedule is going to have to follow a school schedule. I can't burn my comp/vacation time without having to consider how it affects my classwork. I can't take it over the summer, as that's when our crunch time for changing out systems is. As you're not working for a university, the summer issue shouldn't be as bad for you.
Anyway, back to the point -- go and talk to people from the college(s) you're interested in. Some might let you test out of classes (for a fee), and some are more understanding than others when you have to miss a class. Some schools specifically cater to what they call 'adult education'. Look for the ones with a good variety of night classes, if nothing else.
Talk to your professors at the beginning of the semester, and let them know what your situation is. Even just a little 'I might have to run in the middle of the class' is nice warning to give them. Some professors will be understanding, and well, others are complete pricks... but hey, if you find out at the beginning, you can always drop the class, and not have to pay the full price, normally.
Well, okay, I guess 'minimalist's not quite the right word, but they limit the subject length here.
I know in the back of your book, you recommend certain tools and equiptment for cooks, however, many of the people here are apartment dwellers, or might even be in a college dorm with a shared kitchen, where they don't have much space for storage, or they move on a regular basis, making that cast iron set a royal pain.
Besides the obvious needs for a decent pot for boiling pasta/potatoes/whatever, and a good pan for sauteing/pan frying, what would you insist on having around, if you were in that situation?
After all, there are some things that benefit from a slow cooking time. Scrambled eggs are one of the best examples -- a nice, slow cook in a slightly warmed cast iron skillet will give you some of the creamest eggs you've ever had. After watching the smoked salmon episode of Good Eats, I wouldn't be suprised if AB had tried cooking eggs with a hair dryer.
I know that most of the chili cook offs that I've seen, there are no beans allowed.
Oh, and as my mom's placed first a few times at some local contents, I'll give you all a little secret -- alcohol helps to meld the flavors faster, for those times when you don't have a week to let it sit in the fridge. Add a beer into your next 2 gallons of chilli. [As you're putting it in a slow cooker, you might want to back off whatever other liquids you're using, to make sure you cook off the alcohol].
A good tech support person can handle bad users. Yes, sometimes you have to put them on hold for a while 'till they calm down some, but there's normally other tricks that you can use to make sure you don't make things worse.
For instance, making comments about user's sexual encounters when you're working in the bible belt is just right out. [but well, I've seen it done].
NEVER talk down to the customer. I mean, even if it's someone that 500 people have called with before, it's a new problem TO THEM. Yes, you can give them the answer and hang up on them, but if you say 'I just had someone call in with this a few minutes ago-- Why don't you try...' it comes off as more personal, you tell the person that they're not the only one having this problem [ie, they're not an idiot], etc.
Finding people with the right personality for doing helpdesk is normally harder than finding someone who knows the technical information. And well, if you don't have good management, they don't pull people as they're getting burned out, and it makes things harder on everyone. [Getting a reputation as having bad service because of one dumbass pisses you off, and if the customer had last talked to that dumbass, they're more likely to be beligerant when calling you, which can also piss you off, overall in driving down the quality of service which you give, etc.]
Just because the customer doesn't know the answer doesn't mean they're an idiot. They could be a plumber, auto mechanic, doctor, or someone else with specialized knowledge that differed from yours. Personally, if I were going in to get my car fixed, I care about if they know my model of car, I couldn't care less if they knew how to plug in a computer, much less use one. I'm not going to hold it against my dental hygenist that she accidentally infected her system with a virus, but I would be annoyed if she didn't do her job.
That's not to say that there aren't rude people out there who will always get bad service. I had someone call me up when I worked at my university's helpdesk in 1994, and she wanted to know what her monitor's resolution was. [Mind you, we were software support for the computer labs, but our department was "Technical Assistance". After I tried to explain to her that she needed to talk to whomever sold her the computer, she said something to the equivalent of 'I called Technical Assistance, and this is a technical question, so I expect a f**king answer'. I passed her along to my manager, as well, it well, that's what managers are for.
There's then the games where you get to construct objects, not the games themselves, which many people have been discussing. They may be cool, but they're a seperate class of item. Yes, designing your own robot is cool, or designing a car, etc, but it's not the same as 'designing' your own game.
The other category which people have commented on are the build your own level type things. Neverwinter Nights, Quake, Halflife, even back to the days of Doom. Yes, they're nice, but they then require the original program to play, and the editors are developed by people other than the people who wrote the engines.
Personally, I'd suggest to people interested in writing their own games to look at muds. Yes, the majority of them are text based, but there are a few graphical muds out there. Many of the text based engines have been released to the public, and there's a graphical engine, Worldforge, but I have no idea what their current status is.
Anyway, an interesting read from the Slashdot archives:
It's an interesting theory, but there are some benefits of metal that you don't get from rocks. For one, if you import metal, it's all of a known quality. With on site rock, you're taking a significant risk--
Is the rock porous? What sort of loading can it take? How risky is it to mine?
Once you have this sort of information, you can determine the risks of attempting what you're suggesting. There may need to be some sort of a sealing process to prevent permeation, and an ultrasound scan to determine if there are any fissures to be worried about. [which they'd have to constantly re-check while mining, and after every major meteor hit in the area].
I do see it as being a definate possibility for living, however, due to the risks involved [ie, accidentally breaking through to a cave which is connected to the surface, and causing decompression], I would assume that they'd want to set up a preliminary construction camp from which to build the caves.
Depending on what you're attempting to develop, you may be able to develop the code faster using code fragments from other developers. Of course, if those bits of code were GPL'd, you'd be obligated to make your code available.
Depending on the scale of the project, and the odds that the code segments you need already exist, you have to determine how much time savings you'd have by researching previous GPL'd projects over writing it on your own.
Although many companies wish to retain the rights to software you write, there are very few people who don't re-use bits from project to project. [Hell, it'd be downright foolish not to use already written and tested code]. As such, on any programing project I take, although there might be an NDA, I still retain the right to re-use the code in any further projects. Otherwise, I run into the risk that my common code library will be locked down once it's in use in this project, and I'd rather not take the project.
[even if they paid me more than my going rate, I'd be worried about using knowledge that I got from the project towards another project, and getting sued.]
I agree that there are times when I want to open a new window, and there are times when I don't. There are times when a good design might want to a user to have a new window, however, the problem that I see with most designs these days is that they attempt to use Javascript for this.
You DO NOT need Javascript to cause the link to pop up in a new window. If you're one of those fascist bastards who has to control the user's experience (take off the chrome, etc), you might need javascript, but if it's just going to pop up in a new window, use TARGET inside the <A>. You can even specify it inside a <FORM> tag.
This way, when we, as content consumers, wish to open the link in a new window, and we choose 'Open Link In New Window', we don't get blank window with a javascript error at the top, because the source for the javascript isn't in the new window.
As with anything in design, there are right times, and wrong times to use just about anything. Just because there are right times to use Feature X, it doesn't mean that Application Y is that right time.
The most important advise I can give you is to know your audience.
Personally, I'm more interested in how many minor violations that can be handled with this. One of the main problems with quality of life in the DC area is the traffic. With a few cameras in strategic places, the police can respond better to traffic issues, such as accidents, traffic lights out, gridlocking [I hate K street], people parking in the 'no parking 4:30pm-6:30pm' zones (17th NW), or people double parking on both sides of the street, creating a slolom on F St NW, between 17th & 19th.
Oh...and there's a rather large number of buildings which don't require going through metal detectors. The feds may have 'em, but I know the private buildings don't. I haven't been to the Smithsonian or any of the museums, so I don't know they've reacted so far.
If the cameras are used right, I'm all for 'em, and I come into DC every weekday, and some weekends. For those who aren't in the area, and are bitching about the cameras, don't come to DC. The traffic's bad enough without you here.
I mean, when I think 'double duty', I think of it doing two things at once. And well, in the case of a computer, it's heavy, it generates heat, and it may be strong enough to prop something up.
So in that case, I'd have to say that I've used a Wang before to keep pizza warm in the machine room (it was just the right size on top for a large pizza, and conveniently table height).
I've seen towers used as stools, older generation systems (Honeywell, Wang, etc), used as tables, but in those cases, they're normally not still running, or they're running next to nothing, so they might as well not be doing their primary job.
Then we get to what I think of as 'recycling' more than 'double duty', which normally involves keeping the system from being able to do its primary function because it's doing something else. Fishtanks, tables, shelves (IBM 7171 terminal server; Wang) and beer fridges all seem to be fairly common.
(Please note -- there's more than one set of pictures, and they're still a year old...the house may be messy, but it's no where near where it used to be, now that my brother took most of his stuff to college, and isn't using my place for storage. I don't have any recent pictures, so the Honeywell's not there yet, and the reel-to-reel tape drive's buried in crap in the pictures I have)
You have never worked at a university
on
Disaster Recovery?
·
· Score: 2
Most of the system administrators would love to be able to consolidate systems down to a few supported platforms, break services up so we're not supporting 'mainframe-esque' systems running 20+ applications. Unfortunately, the system admins are not what drives the university. Research dollars are. And to get the research dollars, you have to have faculty, and you have to keep them happy, which means installing undocumented software at some faculty member's whim, or keeping a Wang still running so that they can do their word processing. [It does, however, keep pizza warm, so it's not all bad].
Yes, some systems can be consolidated down, upgrade, or otherwise be made more space efficient, but you need to maintain the same OS and similar hardware, or you're looking at significantly increasing your workload due to instability, installation headaches, etc.
Now, there may be some systems that just can't be recovered, but it's not the system admin's job to decide that-- it's management's. The system admin can give advice, but they don't run the university, and if management decides that it's in the best interest to restore 17 year old mainframes that suck down $300k/yr in maintaince contracts and cooling costs, and occupy 1/2 of the space in the machine room, it's their decision. You can either do what they tell you to, or find a new job.
Now, if your management repeatedly doesn't listen to you, and continues to do what you warn them against, you'd probably be happier finding a new job. One of the nice benefits of university jobs is that you can carry over TIAA-CREF to most schools.
Personally, I would recommend first recovering every system possible, and once that's done, and you have everything back and operational again, work on migrating out machines that are harder to maintain and recover. Do not worry about getting rid of systems unless they just can't be recovered. Don't worry about anything else until after the systems are recovered.
PS. Don't put machine rooms in basements. Sewer pipes breaking above the machine room is bad.
I guess I should have stressed flawed logic more than flawed math.
If you're dealing with a true Poisson Distribution, (ie, queuing theory), you have to assume what wait you're willing to deal with. You also assume that things are rarely at capacity, which in our case, was a correct assumption.
In this case, it's odd, as you would never assume that your users are taking up 100% of the bandwidth of their modem at the same time that every other customer is taking up 100% of their bandwidth. [Which is why there's a concept of CIR in frame relay, etc].
You also can't assume that the line to some other ISP is the main cost of business, as it's almost an order of magnitude lower than the cost of PRIs to fill it. [8 PRI per T1 was a good ratio for our users, for which we were paying over $1600/PRI, and $1400 for a T1]. If you assume that you're paying someone $10/hr, 4 work weeks would be $1600. You might be able to get by with that sort of pay for tech support, but not for system and network administrators. [Again, we have economy of scale, as you just can't get 'partial' folks, and you can add to a good system admin with a medicore one, but you don't want to just have the mediocre one, even if they are cheaper]
The larger problem is that we have someone who doesn't fully understanding the problem who thinks they know the whole story, but they don't. Quite simply, an ISP can pull in a hell of a lot more than what the poster assumed, however, they also have to pay out a hell of a lot. To make things worse, it's was a very competitive business, as you had folks who were trying to grow as fast as they could, so they could IPO, or sell out, and make a fast million. Yes, there were CLECs who had tarrifed PRIs for under $500/month, as they were making money from reciprocal charges to the ILEC. Unfortunately, the baby bells fought that, which is why ISP traffic is now exempt from those charges.
Just because you're in a tech field doesn't give you the right to make assumptions about other technical fields, no matter how similar they are. You don't see some pediatrician making claims about how easy it would be to clone people. You don't see cabbies talking about how easy it would be drive in the NASCAR circuit. Just because you work with some version posix OS doesn't mean that you're qualified to install and administer a Sun E10k, but it seems that computer folks overstep their boundries all the time. Don't even get me started about the consultants.
As someone who's worked at a mid-sized ISP, I can tell you that your math is flawed in many ways:
You can get more than 50 modems perl T1. Although folks might be dialed up, they're not filling their pipe the whole time. You're looking at 100-200 "56k" modems per T1. [depending on scale and your exact user base]
15 users per modem is horrible. 6-8 is a much better range. Again, it's based on your user base, however.
ISPs have much bigger charges than the T1. First, you have whatever debts you're paying off for your router and modems. Then you have the recurring charges... Either POTS, Channelized-T1 or PRI. Depending on where you are, and the economy of scale, an ISP could be paying anywhere between $30-80 in recurring costs per month, per incoming line.
Now, based on those numbers, if you've got 115 modems (5 PRI), and you're paying $80 per line, ($9200) and $1600 for the T1. ($10,800 total). You're charging $20/mo, and keep a user/modem ratio of 8, for $18,400.
So, we've got $7600 profit, right? Well, no. There's still business phone lines, loop charges, location rent, utility bills [ie, electricity], ongoing costs of equipment upgrades, etc. So, say you're not paying that much, and you're pulling in $7000 per month (which would be damned high, mind you).
Well, that's $7k/month, or $84k/year. Sure would be sweet, but unfortunately, you probably need some other folks to help you run the place, or you'd have to do all of the tech support, 24x7 network support, billing, accounting, etc, on your own.
ISPs are profitable, but it's a sliding scale... if you upgrade too fast, you pay our more to keep the customers happy, and cut into profits. If you don't upgrade fast enough, you have constant busy signals, and you lose customers, which cuts into profits. You have to be slightly forward thinking (as it might take 2-3 months to get that PRI in from the order date), but you can't be too over enthusiastic.
However, as with any business, you don't _have_ to serve people. If you have a problem customer, you can get rid of 'em. It's perfectly legal, and well, the AUP/TOS just helps to cover your ass. Yes, they might bitch, but when you're paying $80/month for modem line and hardware charges, you've suddenly stopped losing $60/month on that person.
[There are, however, ways to handle the problem customers, but I'd have to classify that as secret, as I still have a vested interest in the ISP]
Is when you have some sort of a help-desk type environment, where the client comes in to talk, and you need to show them something. There was an IBM LCD that would do a sort of a backflip for that purpose:
As anyone who's attempted to use VNC to fix a server while someone's sitting on the system, wrestling for mouse control is a pain in the ass.
I have however, had luck with a two keyboard system [Macintosh], when some friends from Galudet would come by, we'd set two folks up on a system with two keyboards, so they could communicate without having to write everything down, and keep passing paper back and forth.
I remember taking a few flights through National Airport, back in the Metricom days. [Richochet, or however they spelled it]. It does sometimes take a little bit of time in finding power outlets, and in making sure you're near them, but if you have a good eye, you'll find them.
[Simple fact is -- they use 'em for vaccuum cleaners. Some airports have 'em in the floor, with some sort of locking panel over 'em...as we can't carry pocket pliers or pocket knives on us anymore, we can't open those up, but quite a few airports have 'em just as normal circuits along the wall, under some seats.]
The Coolio case was an isolated incident due to miscommunication. He does in fact get permissions (or well, in that case, thinks he has permission), these days. Although that may not have been the case when he was sending stuff to Dr. Demento, it is now.
I have no idea why you would assume that due to one incident, that they were all that way.
Also, Wierd Al has quite a few songs which are not paraodies (completely original, but with the basic influence of a certain style, such as "Stuck in a Closet with Vanna White" or "Albequerque") Also he has a number of polkas which use the lyrics, but not the music from popular songs at the time (to point out just how wierd the lyrics are.)
He's also been turned down to do parodies, and well, you haven't seen any Prince ones out there.
I've seen blackboard mentioned, but that's normally just one part of the puzzle. Blackboard doesn't handle all of the class registration part of the puzzle. There is one, called 'Banner' from SCT, which handles it, but from my personal experience, it's a pain in the ass, as every upgrade to the system requires re-applying your configurations to it. [Which might take multiple man-years to apply for some colleges]
Supposedly, PeopleSoft has a module specifically for educational institutions, but I've never seen it. I also know that there was work being done up at Harvard back in 1996 for basically what you said above. I have no idea what ever became of it.
As for getting students to use the system -- students have a 4 year turnover. You can get an over 95% compliance in 4 years just because they don't know what the old system was. Your problem lies in administration. And you can't have students do this work, due to restrictions by FERPA. It would have to be tightly controled by the Office of the Registrar or the equivalent office.
And as a person who was one of those students making enterprise solutions in the mid 1990s, I'd have to say that student run projects are bound to fail in the long run for larger instutions, due to the lack of documentation, and incorrect dependancies on legacy systems. Although students may make good programmers, major projects need to be led by full time personel who are directly responsible for the project.
(and as for wireless...my university was one of the test beds for Richochet in 1995. Damned nice system for $300/yr at the time... too bad they went under)
Slashdot Mirror
If someone were to 'extend' the way that an LDAP consumer interacts with a supplier, there's an issue. As it is now, every set of LDAP software has its own way to handle direct replication of data.
/etc/passwd normally).
However, the type of information that can be placed into an LDAP directory is handled by the schemas.
In much the same way, a database server doesn't care what your table definitions are to maintain ODBC compliance.
Now, in this case, we have schemas, which much like a table definition, define what sort of information is allowed to be placed into the directory. It may be mail addributes, such as 'rfc822mailalias' and 'mailforwardingaddress', it might be posix authentication information, such as 'loginshell' and 'homedirectory'.
In this case, I have no idea what schema extentions OS X may require, but the standard posix attributes would make sense (ie, what you'd find in
Damn, and people thought I'm strange for making bets on when the movie actually starts. (I'm doing best w/ 12 minutes after the posted time for the movie to start, but I've noticed that sometimes the day the movies come out, there's only about 9-10 minutes of crap).
Oh...and before you make bets, make sure everyone agrees on the clock that you're using. I normally use my cell phone, as keeps better time than wrist watches.
Let's see, there's um....wait, no, that's command line.
/, and configured SIMS to use one large mail partition, without mentioning to us that the SIMS backup program stops when it hits a 2G limit, and that's the only reason we were making our 4hr backup window. [over 300G of data]]
There's um....wait, no, that's SunOne, which is really iPlanet, which was the Sun/Netscape Alliance, which was Netscape...
um.... well, the OS installer doesn't suck that bad...assuming you're working from a jumpstart server, and don't need to work interactively.
[Oh...from experience...you're still better off trying to install it yourself than letting Sun Professional Services do it for you... installed all the applications in
ISPs are value-added resellers, who sell a product based on queuing theory.
What you seem to be confused about is the two parts to the charge, which normally isn't seperated out on DSL and other smaller connections.
There's line, and there's port.
Line is the charge for the connection between you, and the ISP. Sometimes there's markup, sometimes there isn't. There's normally markup if the ISP owns the equipment used to terminate the line. No real queuing here when you're dealing with a form of dedicated circut, which would be basically anything but standard dial-up.
Port is the charge for your data to reach the internet. You always have to deal with queing theory here, or well, everyone gets screwed. You may have a max of 1.5kb download, but is there any sort of a CIR on the line? [what's the speed that they'll guarantee you to get? I'm guessing, nothing, as it's not in your contract with them].
Take the example of the ISP that I used to work for [before they started doing DSL]. We had somewhere near 200 modems per T1. With all 200 people using a 56k modem at the same time, that'd be over 10mb...but a T1's only 1.5mb. In our case you could always assume a certain amount of slack on the lines -- even when a person's connected up, it's not working full tilt, or it's doing something local, like downloading e-mail.
Most ISPs have something to the effect of requiring 'interactive' sessions or something similar, to keep you from sucking down the bandwidth with servers or bots. If you wanted to run a server, upgrade your plan, and get a business plan, or dedicated dial-up, or something similar.
People who try to get stuff for free, and take the ISPs for a ride are the reason that the market crapped out. You have a contract of service with your ISP. If you are using services which are specifically denied by your AUP, you are stealing services from them. It's no different than going to a store and lifting something. Just because it CAN be done, doesn't mean that it should be done.
Oh...and all ISPs should just be banning all incoming traffic on known ports if they have any hopes of staying alive.
We had to have a forced power outage while they were installing power cutover, so they could service the UPSes in our rather aged datacenter.
As we had between 2-4 hours that we were going to be down, as we hadn't gotten approval for a maintenance window in over a year, we moved some disk arrays around, replaced out some old sun "Mass Storage" racks [30" wide cabinet, 28RU, single power] for "Expansion Cabinets" [24" wide, 36U, dual powered].
In the process, we also moved a few systems that were in a rack that was primarily filled with a system going out of production, and cleaned up a signficant chunk of the rats nest of cabling behind these cabinets.
We were thanked by the data center management for making everything look nicer. [Not placing the systems with redundant power supplies in a rack with redundant power supplies, preventing future downtime for systems, creating more room in the machine room, etc, etc, etc] Hell, we weren't even directly thanked for pulling it off with less than 2 weeks notice, and only losing one system in the process [Sparc Storage Array 200, which had been in service for over 6 years with no signficant downtime]
Oh...and even more important than looking nice -- it helps if it actually fits through the door, and doesn't cause the floor to buckle.
Personally, I work for a university, so I get to take classes, and they're rather flexible about my taking classes in the middle of the day, so long as it doesn't affect a production service.
... but hey, if you find out at the beginning, you can always drop the class, and not have to pay the full price, normally.
First off, remember that your schedule is going to have to follow a school schedule. I can't burn my comp/vacation time without having to consider how it affects my classwork. I can't take it over the summer, as that's when our crunch time for changing out systems is. As you're not working for a university, the summer issue shouldn't be as bad for you.
Anyway, back to the point -- go and talk to people from the college(s) you're interested in. Some might let you test out of classes (for a fee), and some are more understanding than others when you have to miss a class. Some schools specifically cater to what they call 'adult education'. Look for the ones with a good variety of night classes, if nothing else.
Talk to your professors at the beginning of the semester, and let them know what your situation is. Even just a little 'I might have to run in the middle of the class' is nice warning to give them. Some professors will be understanding, and well, others are complete pricks
This was an interesting article that I stumbled across earlier today, when looking for PVR software:I'm guessing that rags like Advertising Week would have a similar perspective on things.
Well, okay, I guess 'minimalist's not quite the right word, but they limit the subject length here.
I know in the back of your book, you recommend certain tools and equiptment for cooks, however, many of the people here are apartment dwellers, or might even be in a college dorm with a shared kitchen, where they don't have much space for storage, or they move on a regular basis, making that cast iron set a royal pain.
Besides the obvious needs for a decent pot for boiling pasta/potatoes/whatever, and a good pan for sauteing/pan frying, what would you insist on having around, if you were in that situation?
After all, there are some things that benefit from a slow cooking time. Scrambled eggs are one of the best examples -- a nice, slow cook in a slightly warmed cast iron skillet will give you some of the creamest eggs you've ever had. After watching the smoked salmon episode of Good Eats, I wouldn't be suprised if AB had tried cooking eggs with a hair dryer.
I know that most of the chili cook offs that I've seen, there are no beans allowed.
Oh, and as my mom's placed first a few times at some local contents, I'll give you all a little secret -- alcohol helps to meld the flavors faster, for those times when you don't have a week to let it sit in the fridge. Add a beer into your next 2 gallons of chilli. [As you're putting it in a slow cooker, you might want to back off whatever other liquids you're using, to make sure you cook off the alcohol].
Blah...after a bit of searching, I found the article that I just read last week about the "Signs" trailer by Ant Farm:
L ERS.html?pagewanted=all
http://www.nytimes.com/2002/07/28/magazine/28TRAI
They said that some of the most re-used music comes from "Dave", "Hoffa" and "Dragon: The Bruce Lee Story".
A good tech support person can handle bad users. Yes, sometimes you have to put them on hold for a while 'till they calm down some, but there's normally other tricks that you can use to make sure you don't make things worse.
...' it comes off as more personal, you tell the person that they're not the only one having this problem [ie, they're not an idiot], etc.
For instance, making comments about user's sexual encounters when you're working in the bible belt is just right out. [but well, I've seen it done].
NEVER talk down to the customer. I mean, even if it's someone that 500 people have called with before, it's a new problem TO THEM. Yes, you can give them the answer and hang up on them, but if you say 'I just had someone call in with this a few minutes ago-- Why don't you try
Finding people with the right personality for doing helpdesk is normally harder than finding someone who knows the technical information. And well, if you don't have good management, they don't pull people as they're getting burned out, and it makes things harder on everyone. [Getting a reputation as having bad service because of one dumbass pisses you off, and if the customer had last talked to that dumbass, they're more likely to be beligerant when calling you, which can also piss you off, overall in driving down the quality of service which you give, etc.]
Just because the customer doesn't know the answer doesn't mean they're an idiot. They could be a plumber, auto mechanic, doctor, or someone else with specialized knowledge that differed from yours. Personally, if I were going in to get my car fixed, I care about if they know my model of car, I couldn't care less if they knew how to plug in a computer, much less use one. I'm not going to hold it against my dental hygenist that she accidentally infected her system with a virus, but I would be annoyed if she didn't do her job.
That's not to say that there aren't rude people out there who will always get bad service. I had someone call me up when I worked at my university's helpdesk in 1994, and she wanted to know what her monitor's resolution was. [Mind you, we were software support for the computer labs, but our department was "Technical Assistance". After I tried to explain to her that she needed to talk to whomever sold her the computer, she said something to the equivalent of 'I called Technical Assistance, and this is a technical question, so I expect a f**king answer'. I passed her along to my manager, as well, it well, that's what managers are for.
There's then the games where you get to construct objects, not the games themselves, which many people have been discussing. They may be cool, but they're a seperate class of item. Yes, designing your own robot is cool, or designing a car, etc, but it's not the same as 'designing' your own game.
The other category which people have commented on are the build your own level type things. Neverwinter Nights, Quake, Halflife, even back to the days of Doom. Yes, they're nice, but they then require the original program to play, and the editors are developed by people other than the people who wrote the engines.
Personally, I'd suggest to people interested in writing their own games to look at muds. Yes, the majority of them are text based, but there are a few graphical muds out there. Many of the text based engines have been released to the public, and there's a graphical engine, Worldforge, but I have no idea what their current status is.
Anyway, an interesting read from the Slashdot archives:
It's an interesting theory, but there are some benefits of metal that you don't get from rocks. For one, if you import metal, it's all of a known quality. With on site rock, you're taking a significant risk--
Is the rock porous? What sort of loading can it take? How risky is it to mine?
Once you have this sort of information, you can determine the risks of attempting what you're suggesting. There may need to be some sort of a sealing process to prevent permeation, and an ultrasound scan to determine if there are any fissures to be worried about. [which they'd have to constantly re-check while mining, and after every major meteor hit in the area].
I do see it as being a definate possibility for living, however, due to the risks involved [ie, accidentally breaking through to a cave which is connected to the surface, and causing decompression], I would assume that they'd want to set up a preliminary construction camp from which to build the caves.
Depending on what you're attempting to develop, you may be able to develop the code faster using code fragments from other developers. Of course, if those bits of code were GPL'd, you'd be obligated to make your code available.
Depending on the scale of the project, and the odds that the code segments you need already exist, you have to determine how much time savings you'd have by researching previous GPL'd projects over writing it on your own.
Although many companies wish to retain the rights to software you write, there are very few people who don't re-use bits from project to project. [Hell, it'd be downright foolish not to use already written and tested code]. As such, on any programing project I take, although there might be an NDA, I still retain the right to re-use the code in any further projects. Otherwise, I run into the risk that my common code library will be locked down once it's in use in this project, and I'd rather not take the project.
[even if they paid me more than my going rate, I'd be worried about using knowledge that I got from the project towards another project, and getting sued.]
I agree that there are times when I want to open a new window, and there are times when I don't. There are times when a good design might want to a user to have a new window, however, the problem that I see with most designs these days is that they attempt to use Javascript for this.
You DO NOT need Javascript to cause the link to pop up in a new window. If you're one of those fascist bastards who has to control the user's experience (take off the chrome, etc), you might need javascript, but if it's just going to pop up in a new window, use TARGET inside the <A>. You can even specify it inside a <FORM> tag.
This way, when we, as content consumers, wish to open the link in a new window, and we choose 'Open Link In New Window', we don't get blank window with a javascript error at the top, because the source for the javascript isn't in the new window.
As with anything in design, there are right times, and wrong times to use just about anything. Just because there are right times to use Feature X, it doesn't mean that Application Y is that right time.
The most important advise I can give you is to know your audience.
Personally, I'm more interested in how many minor violations that can be handled with this. One of the main problems with quality of life in the DC area is the traffic. With a few cameras in strategic places, the police can respond better to traffic issues, such as accidents, traffic lights out, gridlocking [I hate K street], people parking in the 'no parking 4:30pm-6:30pm' zones (17th NW), or people double parking on both sides of the street, creating a slolom on F St NW, between 17th & 19th.
Oh...and there's a rather large number of buildings which don't require going through metal detectors. The feds may have 'em, but I know the private buildings don't. I haven't been to the Smithsonian or any of the museums, so I don't know they've reacted so far.
If the cameras are used right, I'm all for 'em, and I come into DC every weekday, and some weekends. For those who aren't in the area, and are bitching about the cameras, don't come to DC. The traffic's bad enough without you here.
I mean, when I think 'double duty', I think of it doing two things at once. And well, in the case of a computer, it's heavy, it generates heat, and it may be strong enough to prop something up.
So in that case, I'd have to say that I've used a Wang before to keep pizza warm in the machine room (it was just the right size on top for a large pizza, and conveniently table height).
I've seen towers used as stools, older generation systems (Honeywell, Wang, etc), used as tables, but in those cases, they're normally not still running, or they're running next to nothing, so they might as well not be doing their primary job.
Then we get to what I think of as 'recycling' more than 'double duty', which normally involves keeping the system from being able to do its primary function because it's doing something else. Fishtanks, tables, shelves (IBM 7171 terminal server; Wang) and beer fridges all seem to be fairly common.
(Please note -- there's more than one set of pictures, and they're still a year old...the house may be messy, but it's no where near where it used to be, now that my brother took most of his stuff to college, and isn't using my place for storage. I don't have any recent pictures, so the Honeywell's not there yet, and the reel-to-reel tape drive's buried in crap in the pictures I have)
Most of the system administrators would love to be able to consolidate systems down to a few supported platforms, break services up so we're not supporting 'mainframe-esque' systems running 20+ applications. Unfortunately, the system admins are not what drives the university. Research dollars are. And to get the research dollars, you have to have faculty, and you have to keep them happy, which means installing undocumented software at some faculty member's whim, or keeping a Wang still running so that they can do their word processing. [It does, however, keep pizza warm, so it's not all bad].
Yes, some systems can be consolidated down, upgrade, or otherwise be made more space efficient, but you need to maintain the same OS and similar hardware, or you're looking at significantly increasing your workload due to instability, installation headaches, etc.
Now, there may be some systems that just can't be recovered, but it's not the system admin's job to decide that-- it's management's. The system admin can give advice, but they don't run the university, and if management decides that it's in the best interest to restore 17 year old mainframes that suck down $300k/yr in maintaince contracts and cooling costs, and occupy 1/2 of the space in the machine room, it's their decision. You can either do what they tell you to, or find a new job.
Now, if your management repeatedly doesn't listen to you, and continues to do what you warn them against, you'd probably be happier finding a new job. One of the nice benefits of university jobs is that you can carry over TIAA-CREF to most schools.
Personally, I would recommend first recovering every system possible, and once that's done, and you have everything back and operational again, work on migrating out machines that are harder to maintain and recover. Do not worry about getting rid of systems unless they just can't be recovered. Don't worry about anything else until after the systems are recovered.
PS. Don't put machine rooms in basements. Sewer pipes breaking above the machine room is bad.
I guess I should have stressed flawed logic more than flawed math.
If you're dealing with a true Poisson Distribution, (ie, queuing theory), you have to assume what wait you're willing to deal with. You also assume that things are rarely at capacity, which in our case, was a correct assumption.
In this case, it's odd, as you would never assume that your users are taking up 100% of the bandwidth of their modem at the same time that every other customer is taking up 100% of their bandwidth. [Which is why there's a concept of CIR in frame relay, etc].
You also can't assume that the line to some other ISP is the main cost of business, as it's almost an order of magnitude lower than the cost of PRIs to fill it. [8 PRI per T1 was a good ratio for our users, for which we were paying over $1600/PRI, and $1400 for a T1]. If you assume that you're paying someone $10/hr, 4 work weeks would be $1600. You might be able to get by with that sort of pay for tech support, but not for system and network administrators. [Again, we have economy of scale, as you just can't get 'partial' folks, and you can add to a good system admin with a medicore one, but you don't want to just have the mediocre one, even if they are cheaper]
The larger problem is that we have someone who doesn't fully understanding the problem who thinks they know the whole story, but they don't. Quite simply, an ISP can pull in a hell of a lot more than what the poster assumed, however, they also have to pay out a hell of a lot. To make things worse, it's was a very competitive business, as you had folks who were trying to grow as fast as they could, so they could IPO, or sell out, and make a fast million. Yes, there were CLECs who had tarrifed PRIs for under $500/month, as they were making money from reciprocal charges to the ILEC. Unfortunately, the baby bells fought that, which is why ISP traffic is now exempt from those charges.
Just because you're in a tech field doesn't give you the right to make assumptions about other technical fields, no matter how similar they are. You don't see some pediatrician making claims about how easy it would be to clone people. You don't see cabbies talking about how easy it would be drive in the NASCAR circuit. Just because you work with some version posix OS doesn't mean that you're qualified to install and administer a Sun E10k, but it seems that computer folks overstep their boundries all the time. Don't even get me started about the consultants.
- You can get more than 50 modems perl T1. Although folks might be dialed up, they're not filling their pipe the whole time. You're looking at 100-200 "56k" modems per T1. [depending on scale and your exact user base]
- 15 users per modem is horrible. 6-8 is a much better range. Again, it's based on your user base, however.
- ISPs have much bigger charges than the T1. First, you have whatever debts you're paying off for your router and modems. Then you have the recurring charges... Either POTS, Channelized-T1 or PRI. Depending on where you are, and the economy of scale, an ISP could be paying anywhere between $30-80 in recurring costs per month, per incoming line.
Now, based on those numbers, if you've got 115 modems (5 PRI), and you're paying $80 per line, ($9200) and $1600 for the T1. ($10,800 total). You're charging $20/mo, and keep a user/modem ratio of 8, for $18,400.So, we've got $7600 profit, right? Well, no. There's still business phone lines, loop charges, location rent, utility bills [ie, electricity], ongoing costs of equipment upgrades, etc. So, say you're not paying that much, and you're pulling in $7000 per month (which would be damned high, mind you).
Well, that's $7k/month, or $84k/year. Sure would be sweet, but unfortunately, you probably need some other folks to help you run the place, or you'd have to do all of the tech support, 24x7 network support, billing, accounting, etc, on your own.
ISPs are profitable, but it's a sliding scale... if you upgrade too fast, you pay our more to keep the customers happy, and cut into profits. If you don't upgrade fast enough, you have constant busy signals, and you lose customers, which cuts into profits. You have to be slightly forward thinking (as it might take 2-3 months to get that PRI in from the order date), but you can't be too over enthusiastic.
However, as with any business, you don't _have_ to serve people. If you have a problem customer, you can get rid of 'em. It's perfectly legal, and well, the AUP/TOS just helps to cover your ass. Yes, they might bitch, but when you're paying $80/month for modem line and hardware charges, you've suddenly stopped losing $60/month on that person.
[There are, however, ways to handle the problem customers, but I'd have to classify that as secret, as I still have a vested interest in the ISP]
I have however, had luck with a two keyboard system [Macintosh], when some friends from Galudet would come by, we'd set two folks up on a system with two keyboards, so they could communicate without having to write everything down, and keep passing paper back and forth.
I remember taking a few flights through National Airport, back in the Metricom days. [Richochet, or however they spelled it]. It does sometimes take a little bit of time in finding power outlets, and in making sure you're near them, but if you have a good eye, you'll find them.
[Simple fact is -- they use 'em for vaccuum cleaners. Some airports have 'em in the floor, with some sort of locking panel over 'em...as we can't carry pocket pliers or pocket knives on us anymore, we can't open those up, but quite a few airports have 'em just as normal circuits along the wall, under some seats.]
Also, Wierd Al has quite a few songs which are not paraodies (completely original, but with the basic influence of a certain style, such as "Stuck in a Closet with Vanna White" or "Albequerque") Also he has a number of polkas which use the lyrics, but not the music from popular songs at the time (to point out just how wierd the lyrics are.)
He's also been turned down to do parodies, and well, you haven't seen any Prince ones out there.
I've seen blackboard mentioned, but that's normally just one part of the puzzle. Blackboard doesn't handle all of the class registration part of the puzzle. There is one, called 'Banner' from SCT, which handles it, but from my personal experience, it's a pain in the ass, as every upgrade to the system requires re-applying your configurations to it. [Which might take multiple man-years to apply for some colleges]
Supposedly, PeopleSoft has a module specifically for educational institutions, but I've never seen it. I also know that there was work being done up at Harvard back in 1996 for basically what you said above. I have no idea what ever became of it.
As for getting students to use the system -- students have a 4 year turnover. You can get an over 95% compliance in 4 years just because they don't know what the old system was. Your problem lies in administration. And you can't have students do this work, due to restrictions by FERPA. It would have to be tightly controled by the Office of the Registrar or the equivalent office.
And as a person who was one of those students making enterprise solutions in the mid 1990s, I'd have to say that student run projects are bound to fail in the long run for larger instutions, due to the lack of documentation, and incorrect dependancies on legacy systems. Although students may make good programmers, major projects need to be led by full time personel who are directly responsible for the project.
(and as for wireless...my university was one of the test beds for Richochet in 1995. Damned nice system for $300/yr at the time... too bad they went under)