Ahhh, as far as I know, DNS is still distributed under this model. Google is not hosting DNS, they are providing a look-up server for DNS, just like your current ISP does (usually set when you get your IP address via DHCP). The root servers still exist, I'm still serving my domain's DNS, HP is still serving their DNS,...
Google is providing an alternative to the ISP's servers for look-up, not hosting.
I've got development, ops, part of product planning, and the usual host of others. I 'help' with QA and support which are not mine. AND I do admin work.
Our products, at one point, spanned about a dozen operating systems. Between test beds, build boxes, desktops, shared development boxes, infrastructure, etc, we have 60+ systems. I need a little help:).
More to do with flags to ask questions about. Usually it is an issue with a product or a luser. Flags things I might not hear about. As an example, "Person A had weird PC behavior, they thought is might be a virus because ???? so I ran a bunch of scans, everything was clean, and the issue 'went away'." Since nothing was found and the behavior resolved itself, it might not make an issues list. It will make the "I spent two hours working on A's PC" timesink list, which might make me ask "what happened". If it happens a few times, I MIGHT want to dig deeper as to surfing behavior, equipment replacement, software configs, process/policy changes,...
I trust my admin, if I don't it is definitely time to get a new one. Like all people on the daily firing line, he does not always see the policy issues, long term process issues, or in some cases the 'recurring' issues related to what might NOT be independent events.
Basic premise is that sometimes the guy in the back of the room WITHOUT the extinguisher in his hand can see a better way to put the fire out for good. Also sometimes people live with issues because they have insufficient title/clout/juice to change things. Having a guy in the back looking things over can help, especially if he HAS the title/clout/juice to potentially make a difference.
Sure the same could be said for 10 minutes vs. an hour if there are enough 10 minute issues. The time value is more an issue of payback on my time, avoiding micromanagement,...
Alrighty, I AM a CTO of a 20 person company with a single admin and here is what I'm interested in.
1 - problems and their resolution
2 - potential issues
3 - time sinks.
So I get info on: What broke last week and how did we fix it: a list of hardware software outages, their root causes, the fix applied, whether that fix is a long term of short term fix, if short term, a recommendation for a long term solution
Issues that my admin sees as 'near term' problems (2 months): list of systems low on resources (disk, cpu, ram,...), applications that have repeat issues, upgrades that are due and are non trivial (potential downtime, critical app where the 'upgrade gone wrong' may lead to down time,...), etc. This includes a list of any planned downtime and a description of the planned downtime (including 'the plan'/timetable of events) so I can remind or co-ordinate with others.
Issues that my admin sees as 'mid term' problems (12 months): list of systems due for replacement, applications/OSes that are near end of life, need for additional hardware (network switches, firewall upgrades,...), etc
Any single issue that he spent more than an hour on or anything he is repeatedly spending time on, those are my definition of time sink.
Why am I interested in those specific items:
Items in category 1 are apt to come up in conversation with my boss. They are also items I need to monitor to ensure that the systems, applications, and yes the admin staff, are not causing the company headaches.
Items in categories 2 and 3 fall into planning and budget issues that I need to plan for, or co-ordinate with others.
Category 3 also allows me to eventually understand that application A or staff member B or 'department' C are killing us and I need to find a better way for the company to work. It also allows me a better understanding of whether the week is an anomaly or if I need additional admin staff or training.
None of this is in a rigid format, so no I can't forward you a template:). It is currently done via office visits/conversations, emails, and hallway conversations. That is working and I see no need for a more rigid structure unless we start to have communications issues. When we do, I'll setup a more formal status reporting system (currently, if it ain't broke...).
Bottom line, in a small company, single admin case where that admin reports to the CTO, the CTO is effectively the systems/IT manager as well as the development manager, the CTO or corporate level planner, and the executive level consultant/evangelist on IT matters to the CEO and CFO. I do NOT necessarily expect the admin to be an IT manager, being an admin is frequently hard enough. However, that 'department' is not my only concern so to some extent the admin needs to summarize stuff and not ship me logs/raw data, I have too many hats.
From a best practices approach, we use MANY passwords (probably not quite unique per system but close). We use a tool to store the passwords (encrypted). The tool shows all passwords as ***** until you specifically select one. So the 'camera' would only see the 'current' password on the screen not the entire list.
Of course, if it focused on my keyboard while I'm typing the master password to get to the list...
Wrong. All the passwords to my network ARE written down. They are maintained in an encrypted file. I can find any password on my net by remembering a single password, the password to the encrypted file. Before you ask, the file is AES encrypted with a strong password. We're not talking a password protected zip file or spreadsheet.
Lot's of software exists to handle just this issue, use it.
IMHO, the biggest problems with any computer deployment in our K-12 classrooms are always support and training. If a school district adopts Linux and open source then who is going to be the admin in charge of updates, patches, server, network, and desktop maintenance, etc? Competent Linux admins are harder to find than people with at least basic knowledge of Mac and Windows and are likely to cost more too. So unless someone within the district, who will not be any worse off for saying no, wants to step up and take on the task of learning to be a Linux admin who is going to manage the whole affair? Also, how many teachers know how to use Linux or are willing to invest the time required to learn? After all, they cannot teach their students that which they themselves do not know. These are not insubstantial difficulties.
So don't use Linux. What's wrong with Gimp/Open Office/... for Windows? Most major mainstream applications that are OSS have a Windows version.
Pretty much ALL schools could use Open Office, GIMP, VLC, Firefox, Dia, Inkscape, Moodle, Apache if they require that type of software (all have Linux, and Windows version, most have OS/X). No one says OSS only runs on Linux.
Schools that want to run Linux can run Linux, schools that want Windows can run windows versions. Applications can be introduced one at a time to allow training ramp up. What's the issue?
Point was, original stats are based on papers pulled after publication. They get pulled when someone with credentials (peer) finds an issue. If 'hey I'm Joe sixpack' calls and says "I think there's a mistake" it doesn't usually get pulled. So since the LHC stuff is peer reviewed, joe sixpack reviewed, and reviewed and commented on by just about everyone AND it still hasn't been pulled...
It must have passed the 'stats test'.
And Yes, peer reviewed ain't what it used to be (probably never was), but that's not the point.
The LHC paper has been 'published'. It has been peer reviewed up the butt. It has not been withdrawn. It obviously then falls into the 'other' 999/1000. Like slashdot is fond of saying: there is nothing to see here, move along.
We've found that most employee's 'you are on good terms with' answer questions even after they leave.:).
Basically, the remove 'privs to ensure everything keeps on ticking' you state is the same as 'remove privs and send home' to us. If we are on good terms and we sent you home early, we aren't shy about calling/emailing if we get desperate.
Again, the difference we see is: send you home early with pay and we get desperate, more likely to help. Remove privs and make you come sit here anyway THEN we get desperate, less likely to help.
Yup, there is a risk you'll tell us to 'go fish' but it hasn't happened yet (in the whole couple of times we've had to call).
Now if your issue is: remove privs and keep them coming in for a couple days and if something breaks, have them fix it. We are in agreement, sort of. Difference is we don't remove privs:). We might delete/disable your personal acct on some critical servers for the last day or two you work (to check for stuff running as you that should be some other role/user) but we haven't changed root/admin password and we haven't tampered with your 'personal' computer. So in my opinion we haven't really removed your privs (you still have 'root', can't get too many more privs). You can go to a console and do whatever you want or have anyone in building logon as 'joe user' then escalate priv via root password (su is your friend on unix/linux and the network\administrator windoze acct still exists on the network). So that all can occur while you are still coming to work, if essential. Usually it is not an issue. we frequently look at your account stuff (visual 'oh crap' inspection), kill your acct in the AM and send you home at lunch with a 'call you if we need you' type deal.
OK, I guess I'll de-cloak and jump in. First background, I'm a CTO with a small software company. By small I'm talking about 30 bodies total. I mention this only to provide some data on my experience/viewpoint (CTO, not developer) and my environment (dev shop, small company) and even with all that said, remember YMMV.
OK, our policy falls in to two categories/buckets: 1 - your privs are removed and you are sent home with pay for the notice period, goodbye don't come in to work. 2- you keep all your privs and you continue to work, thanks for staying during the notice period.
Nothing else makes sense to us. Removing your privs and having you come in just creates a distraction while you talk to other staff, not useful to us.
As to whether you fall into bucket 1 or 2 is the result of conversations among management. Any doubt that you will play nice - goto bucket 1 immediately. Any doubt that you are really needed to complete work - goto bucket 1 immediately. If you both can and will contribute to the project and we do not expect any issues with you working during your notice (poaching employees, causing trouble, etc.) then go to bucket 2.
We have had people we assigned to bucket 1 that were great employees and I'd like to keep. They were not really needed for the project and we sent them home as sort of a last 'paid vacation' from us. No ill will, I'd hire them again. We've also sent people home and taken a hit on the project as the distraction, productivity, or trust factor outweighed the usefulness factor.
Removing your privileges and still having you come in makes absolutely no sense to me. Seems to be the worst of both worlds, you can't really be productive and the low work load can cause you to create distractions for other staff. I just do not get why they want to do that.
So how is knowing that he was talking about an application on DOS 2.0 in 1989 and that the question arose in 1990 going to help now. Or differently, assume it is on windows or linux now and it is in ASP or PHP; now provide information using that that helps him make a similar decision in 2030 when neither Windows nor Linux exists any more, hell the Internet as we know it will not exist. The point is "How do I make the general decision"/"What metric do I need to measure/compare in future decisions", NOT what is the recommendation for today. In future decisions he may not (nay, will not) be using the same environment as today.
Asking about the specific environment and providing specific information related to that environment doesn't really answer his question. It does make "our" job easier as we can all focus on the smaller specific issue and not have to deal with the general complex case. The "question" is what is the general case metric to determine obsolescent platforms. Or a better one would be a methodology that provides a numeric value rating the TCO of an architecture/platform. The question, of course, has no one answer and a general solution is not readily available.
What he will find is, the more general/vague the question, the more general/vague the answer has to be. OR the more the answer resolves only a discrete instance. We can all guess at a platform and provide data. If by "Platform" we mean web based multi-user application or client server application and we restrict our answers to that, we will get some form of response for that discrete 'instance'. We can further restircit to a language, ASP/PHP and provide a more specific answer. That answer will be useful until 20xx when that platform/language 'dies'.
We can try to answer the general question but the answer gets pretty complex/vague/academic. E.g., For any system/platform/entity we can decide a set of properties that are important, say flexibility, availability of resources (developers), etc. and assign weights (e.g., web sites that are relatively static need little implementation flexibility and have low resource consumption due to low change rates). We can add other variables we think important (e.g., lost opportunity costs) and add weights for those. All this continues until we have the massive untested model. The equation of which would contain numerous weighting factors and if written on paper would probably be massive enough to kill a cat (or not).
Now we can continue to run down this rabbit hole but I expect that little of practical value will emerge the other end. The key to the data the submitter wants is, as usual, in asking the correct question. General open ended questions are not known for generating highly useful, reusable, succinct answers. Highly specific questions are more apt to generate highly useful and possibly succinct answers but they frequently leave much to be desired in reusability. The best we can do is state this obvious issue to the original poster and suggest they rethink the issue and pose a question that strikes a correct balance for the information the submitter needs. I personally intend to conserve what few brain cells I have left and move on to other topics.
No he it NOT asking whether he should drop his platform. He is asking the more general question, how do you know when ANY platform has reached it 'drop' time'. You want to feed him a fish, he is asking to be taught to fish, big difference.
That breaks other stuff. The concept of +infinity -infinity is very useful. Otherwise series like x+2x+3x+4x+... converge to the same limit/value as x-2x-3x-4x-... as x->infinity. Kinda like saying all infinities are created equal which breaks lots of advanced math that relies on cantor set issues.
I'm all for simplifications but too simplified also causes issues. A long time ago my son was learning about sqrt. He asked the teacher about sqrt of negative numbers. The reply was "you can't take the sqrt of a negative". First, being a bright child, he ask me when he got home. His statement was, the teacher said you can't do sqrt of negative numbers, that doesn't seem right, math is usually more consistent than that, is it true? A short discussion of complex and imaginary numbers followed. Later in the year the teacher taught imaginary numbers.
The net result of this education was: my son learned to never assume the teacher was telling the whole truth (his statement). That whenever the teacher said something couldn't be done or was wrong it MIGHT just be wrong until the teacher decided to talk about it later. The teacher's attempt at simplification (in the short term) led to a lesson in the fact the teacher oversimplifies to avoid discussion and could not be trusted to tell the 'whole truth'. Not EXACTLY the lesson the teacher hoped for by oversimplifing (I assume).
So while simplification is good, oversimplification to make a point is not always good (and sometimes makes a different point).
I admit to not having read EVERY comment in this thread, so it is possible this is a dup but...
The guy is wrong. The reason divide by zero is undefined is simple:
1/-1... 1/-.5... 1/-.1... Is an stream of negative numbers increasing in absolute value e.g., -1,... -2,... -10,... So as you take the limit of division buy zero from the negative side it goes to negative infinity. Now try:
1/1... 1/.5... 1/.1... and get a stream of positive numbers approaching positive infinity.
Try a graph.
Now explain how a single number/value/nullity can represent BOTH positive and negative infinity. Maybe if I create a function that approachs positive 5 from one direction and -5 from the other direction I can coin the 'fivish' as the solution and get my 15 minutes of fame as well.
Ah well, America seems to own crappy science these days I guess the Brits are entitled to crappy math:).
Think in-city clustered 12 story apt buildings. A simple building, 1 apt per floor, say a 1000 sq ft apt. So we have 12 'stacked houses' that have a single 'shared' 1000 sq ft roof. So 1000 sq ft of roof needs to support 12 'houses' of energy. Unlikely. THAT'S why the grid comes in. In clustered/dense city dwelling environments, roof top energy generation will be insufficient to cover the usage. We'll always need a 'grid' in that environment. How we fuel/supply the grid may be debatable but the existance is required due to the space/environmental constraints.
Oh, I agree that the smaller the blocks the easier it is to 'automagically' prove the code. I'll also agree that 20+ years of dev experience have taught me that small blocks are better even if we do not formally prove anything. Small blocks are more likely to be 'visually verified' by the developer and the follow-on developers, reducing bugs. If you can't hold the whole thing in your head, it is more likely to have issues. Compartmentalization is GOOD.
But more code is more code, if you do not 'prove' the implementation. If compartmentalization reduces the errors per 1000 lines of code from X to Y, say 40% of the original value, you still can not claim a 60% reduction in bugs as the lines of code numbers are different in the different implementations.
To pull numbers out of the air... If you can get a 30% reduction in errors per 1000 lines via a method but you need 2 times the lines you net MORE errors. Assume a statistical 10 errors per 1000 lines and a project of 10,000 lines, that's a likely 100 errors. Now the improved methodology statistically has only 7 errors per 1000 lines (30% better) but the project requires 20,000 lines yielding 140 errors. Now 10,000 of those lines MIGHT be provably correct due to structure, but if you DON'T prove them you have to count them (I agree if you truly prove them you can remove them from the estimate, yielding 70 likely errors).
As I said: In the real world people aren't proving the code, they are claiming it COULD be proven. BZZZZT, thanks for playing. You now have to count the extra code at the lower statistical error rate AND you have to pay the performance penalty. Sometimes the lower rate and the additional code does NOT yield a lower predicted error count so why do it. Sometimes it does yield a lower predicted error count, at a performance hit, yielding a real world trade-off.
ALSO the compartmentalization frequently yields a 'lower exposure' per bug. In a mono kernel exposure is likely all of ring 0. In a microkernel an exposure is not always ring 0, sometimes ring 1 or 2, yielding a POTENTIALLY lower damage/risk. Whether it is ACTUALLY a lower risk depends on how the box/OS is used.
Basically, the whole threat model changes. This leaves you with an apples and pears comparison, not apples to apples. The trade-off needs to understood and an intelligent decision needs to be made. Consequently stating that a general purpose computing platform delivered with a microkernel that has lower performance is better than a monolithic kernel that has better performance is a useless statement (apples to pears).
If you understand the trade-off and the application environment you are applying it to, you can decide if you prefer the shiny apple or the shiny pear. It is NOT a choice between the shiny apple and the wormy apple.
If you want a shiny apple vs. a wormy apple, give me a microkernel environment with better performance than the mono kernel that has less lines of code and a lower statistical error count than the mono kernel and I'll be happy to make the obvious/easy choice.
Actually, I am not interested in proving the kernal correct, no one is in the real world (imo). I am interested in being assured that my PROCESSING is correct. While the micro kernel may be smaller (not equal to simpler) I still need to then look at all the services that the mono kernel supplied that are external to the micro kernel.
In all cases where a general computing OS is used the services list is probably the same (custom kernels are a different matter and a micro may strip down better, or not).
Yes you get SOME advantage in security as MAYBE less things run at ring 0 but still lots of stuff will run at elevated priviledge (ring 0 or 1 not 3) and you have more complexity.
In my opinion performance is one thing you care about, maybe less than security but you care. If the end-to-end security issues are even roughly the same between mono and micro kernels then performance trade-offs start to weigh in more (if I can have 1% better security at a 50% performance loss, do I want the trade? Where is the swap point, 5% security and 25% performance, 10 to 10, 50% better security at a 1% performance hit?). Oh, and yes the brighter reader is correct, it depends on my usage, does is not?
Micro kernel OSs are more complex end-to-end. More complex typically implies more code, more code typically implies more potential bugs. However, microkernels are more compartmentalized, better compartmentalization improves security (broken part A does not imply access to part B). So the intended usage of the kernel impacts its overall end-to-end security 'rating' for the application in comparison to a mono kernel.
A system that is only used by a single user, no network access, physical access restricted, user always runs a root, is probably as secure in either case (and a damn rare bird). A system that is naked on the net running 50 different publicly accessable services (web, DNS, mail, IM,...) is a completely different animal. No one size fits all (shocking I know).
Life is full of trade-offs, this is just another one.
This is the age old argument, 'give a man a fish' or 'teach a man to fish'. Yes we can build roads/power plants/drugs for them, OR we can improve education so they can eventually build their own 'roads'. You appear to be of the 'give a man a fish' school while the media lab is of the 'teach a man to fish' school. You are both right and you are both wrong.
Solving the short term at the expense of the long term, is usually considered bad public policy (but usual public policy). Concentrating on the long term while people suffer is considered bad humanitarian policy (but watching people suffer is usual individual policy). I doubt we'll solve this debate here.
Ahhh, did anyone check the math. It looks a bit off. First he uses a 1.15 multiplier to account for 'other costs' THEN adds it to the loan value (i.e., interest oriented). If you read the endnote that is based on the fact that loans are for 115% of the value (payoff on old car?). How is that a legit 'cost' of the new hybrid car?
Second he is using the full cost of the hybrid. He is assuming that you dump a perfectly good car and buy a hybrid, NOT that you are bright enough to buy a hybrid when it is time to buy something. That is, he is assuming it is the full cost, not the incremental cost of the hybrid. While that MAY be a correct financial analysis, it is unlikely to be a real world analysis (IMO).
If I want a $22K hybrid and my other choice is a $18K car/SUV at 25MPG, then the 'additional capital expense' is $4K NOT $22K. $4K * 1.15 (assuming I use his magic math) is $4.6K incremental cost at 5.25% over 60 months that's about $88/mo in payment. Given the gas savings and higher trade in allowance, the case for a hybrid may be closer than he paints. Of course that assumes the competition for your car dollar is an SUV at 25 MPG if it is a small car at $15K and 30MPG then the hybrid case is less good.
The real issue is during a "I'm going to buy a new car, what will it be" purchase period. It is fair to deal with incremental costs and incremental improvements in gas mileage/trade-in value. As I read it, the article assumes a 'forced trade' at full cost, not incremental costs. I'm not sure that is a fair comparison.
Think easily distributable E-text books and CBT programs. In some areas a teacher comes way after a doctor or field irrigation (not saying it should, but it does). A cheap laptop with hand crank power that is full of e-books on first aid, farming, and basic 3-Rs education is useful as a community resource. (Think library and basic training) It is a one time cost not a recurring cost (OK, recurring every N years if you take care of it). A teacher with a roof, school, and materials is a recurring cost to the community.
Is the approach practical? I do not know. Probably depends a lot on the PC survival rate and quality of material available. I'm sure there are parts of this planet where a PC with a 5 year life at $100 ($20/yr) is a viable alternative to a paper based reference library and the one day a month a teacher might wander through.
If not, it is viable in the US where the average high school student in my area totes 20+ pounds of books between classes. Move to e-books and this low price reader and save the storage, transportation, distribution, and student agony of all that paper hauling.
Potentially 4 times the size folded up neatly and in my pocket so my old 'boomer' eyes can see the type. Think a foldable e-book that I can load with a couple of novels, carry in my pocket, and read on the train.
Add a tiny bit more cpu, an audio out jack, and an MP3 player. Now you have an e-book with MP3 player for the commuter.
Or take a PocketPc/Palm/PDA and add a mono video out jack for one of these and you have a decent e-book screen for a PDA.
Or use the PDA size device, add this screen, add Linux, add a wired and wireless ethernet device, add snort/tcpdump/... and you have a large screen (unfolded) network diag device with PDA keyboard (add USB jack for USB keyboard possibly available at destination).
Bottom line, I prefer a backlit color LCD too. I also prefer long battery life (days not hours), and larger type for my older eyes. Life is full of trade offs. I'll give up the color screen for more battery life and larger print in a commuter device or network diag device.
It is not a laptop and will not replace a laptop in today's world until at least 8-16 bit color is available. Stop thinking laptop and start thinking large screen low weight appliance with longer battery life.
(Note: think a mono laptop (OK 2 bit greyscale) that had old mono PDA battery life (days/weeks), weighed less than 8 ounces, was PDA size with 'fold out' keyboard and wireless would probably have some market share as an e-book/light net surfing (text)/commuter appliance.)
(Note: I like Linux, I run Linux, the company runs Linux in the infrastructure. The following is a business discussion NOT a personal preference.)
OK, go to Dell chose a Linux PC and note the price. Now go to a windows XP Home based PC and 'upgrade' the config to match the Linux PC from a speed/memory/disk/video perspective. Last time I did this the price difference was ~$40 on an $800 PC (5%). Remember that while XP is $200 retail it is ~$30-50 OEM (hey $40 times a million PC per year per 'Dell' adds up).
On Walmart.com I see a celeron D 2.8Ghz 256meg 80G drive Win XP Pro for $407. A celeron D 2.8 Ghz, 128 Meg, 40G drive, no OS is $78 cheaper (note smaller drive and less ram on No OS and the Win box is Win PRO not home). Walmart does not really do support (manufacturer support) so it is less likely to mark up for support costs (the manufacturer to walmart costs MAY already have a markup for it).
So the 'manufacturer' has double the OS test/certification costs to reduce his retail price ~$50 or ~5%. For this effort he gets entry into a small market space. THAT IS THE MAJOR ISSUE. If I have to test on multiple OSes and support multiple OSes I need a reasonable return on that cost. Right now it is not there.
The cost of training the support staff to support Linux doubles or triples my support training costs. I now have N sets of drivers to track and make available on my support web site. I have extra sales training (I want to use Linux and I want to do X, what PC model and software should I buy).
The expectation as has been stated here a zillion times is the PC will 'cost less'. So for my ~$50 in price difference I MIGHT sell a few PCs to the currently small Linux market place. For that I have to control my support costs and sales costs to market two different OS systems in order to keep a profit. (note: assumption, the $50 difference in price is roughly what the OEM cost of Windows is, does anyone know the OEM pricing for large retailers like Dell?).
That's on an entire PC. For a network card manufacturer selling product at ~$20 retail his 'profit' is probably a couple bucks per card, why should he assume the extra cost burden of doing driver dev and testing twice and supporting them?
I can 'almost' double my training and software tracking/'driver dev' costs and support the whole market or get ~90% of the market at half that cost.
Please explain again why the business guys are not making a good business decision?
Slashdot Mirror
Airport + plane ride, Kids in car, field work, ...
Ahhh, as far as I know, DNS is still distributed under this model. Google is not hosting DNS, they are providing a look-up server for DNS, just like your current ISP does (usually set when you get your IP address via DHCP). The root servers still exist, I'm still serving my domain's DNS, HP is still serving their DNS, ...
Google is providing an alternative to the ISP's servers for look-up, not hosting.
I've got development, ops, part of product planning, and the usual host of others. I 'help' with QA and support which are not mine. AND I do admin work.
Our products, at one point, spanned about a dozen operating systems. Between test beds, build boxes, desktops, shared development boxes, infrastructure, etc, we have 60+ systems. I need a little help :).
Nope, nothing to do with his performance.
More to do with flags to ask questions about. Usually it is an issue with a product or a luser. Flags things I might not hear about. As an example, "Person A had weird PC behavior, they thought is might be a virus because ???? so I ran a bunch of scans, everything was clean, and the issue 'went away'." Since nothing was found and the behavior resolved itself, it might not make an issues list. It will make the "I spent two hours working on A's PC" timesink list, which might make me ask "what happened". If it happens a few times, I MIGHT want to dig deeper as to surfing behavior, equipment replacement, software configs, process/policy changes, ...
I trust my admin, if I don't it is definitely time to get a new one. Like all people on the daily firing line, he does not always see the policy issues, long term process issues, or in some cases the 'recurring' issues related to what might NOT be independent events.
Basic premise is that sometimes the guy in the back of the room WITHOUT the extinguisher in his hand can see a better way to put the fire out for good. Also sometimes people live with issues because they have insufficient title/clout/juice to change things. Having a guy in the back looking things over can help, especially if he HAS the title/clout/juice to potentially make a difference.
Sure the same could be said for 10 minutes vs. an hour if there are enough 10 minute issues. The time value is more an issue of payback on my time, avoiding micromanagement, ...
Alrighty, I AM a CTO of a 20 person company with a single admin and here is what I'm interested in.
1 - problems and their resolution
2 - potential issues
3 - time sinks.
So I get info on:
What broke last week and how did we fix it: a list of hardware software outages, their root causes, the fix applied, whether that fix is a long term of short term fix, if short term, a recommendation for a long term solution
Issues that my admin sees as 'near term' problems (2 months): list of systems low on resources (disk, cpu, ram, ...), applications that have repeat issues, upgrades that are due and are non trivial (potential downtime, critical app where the 'upgrade gone wrong' may lead to down time, ...), etc. This includes a list of any planned downtime and a description of the planned downtime (including 'the plan'/timetable of events) so I can remind or co-ordinate with others.
Issues that my admin sees as 'mid term' problems (12 months): list of systems due for replacement, applications/OSes that are near end of life, need for additional hardware (network switches, firewall upgrades, ...), etc
Any single issue that he spent more than an hour on or anything he is repeatedly spending time on, those are my definition of time sink.
Why am I interested in those specific items:
Items in category 1 are apt to come up in conversation with my boss. They are also items I need to monitor to ensure that the systems, applications, and yes the admin staff, are not causing the company headaches.
Items in categories 2 and 3 fall into planning and budget issues that I need to plan for, or co-ordinate with others.
Category 3 also allows me to eventually understand that application A or staff member B or 'department' C are killing us and I need to find a better way for the company to work. It also allows me a better understanding of whether the week is an anomaly or if I need additional admin staff or training.
None of this is in a rigid format, so no I can't forward you a template :). It is currently done via office visits/conversations, emails, and hallway conversations. That is working and I see no need for a more rigid structure unless we start to have communications issues. When we do, I'll setup a more formal status reporting system (currently, if it ain't broke ...).
Bottom line, in a small company, single admin case where that admin reports to the CTO, the CTO is effectively the systems/IT manager as well as the development manager, the CTO or corporate level planner, and the executive level consultant/evangelist on IT matters to the CEO and CFO. I do NOT necessarily expect the admin to be an IT manager, being an admin is frequently hard enough. However, that 'department' is not my only concern so to some extent the admin needs to summarize stuff and not ship me logs/raw data, I have too many hats.
Does that help?
From a best practices approach, we use MANY passwords (probably not quite unique per system but close). We use a tool to store the passwords (encrypted). The tool shows all passwords as ***** until you specifically select one. So the 'camera' would only see the 'current' password on the screen not the entire list.
Of course, if it focused on my keyboard while I'm typing the master password to get to the list ...
Wrong. All the passwords to my network ARE written down. They are maintained in an encrypted file. I can find any password on my net by remembering a single password, the password to the encrypted file. Before you ask, the file is AES encrypted with a strong password. We're not talking a password protected zip file or spreadsheet.
Lot's of software exists to handle just this issue, use it.
IMHO, the biggest problems with any computer deployment in our K-12 classrooms are always support and training. If a school district adopts Linux and open source then who is going to be the admin in charge of updates, patches, server, network, and desktop maintenance, etc? Competent Linux admins are harder to find than people with at least basic knowledge of Mac and Windows and are likely to cost more too. So unless someone within the district, who will not be any worse off for saying no, wants to step up and take on the task of learning to be a Linux admin who is going to manage the whole affair? Also, how many teachers know how to use Linux or are willing to invest the time required to learn? After all, they cannot teach their students that which they themselves do not know. These are not insubstantial difficulties.
So don't use Linux. What's wrong with Gimp/Open Office/... for Windows? Most major mainstream applications that are OSS have a Windows version.
Pretty much ALL schools could use Open Office, GIMP, VLC, Firefox, Dia, Inkscape, Moodle, Apache if they require that type of software (all have Linux, and Windows version, most have OS/X). No one says OSS only runs on Linux.
Schools that want to run Linux can run Linux, schools that want Windows can run windows versions. Applications can be introduced one at a time to allow training ramp up. What's the issue?
Point was, original stats are based on papers pulled after publication. They get pulled when someone with credentials (peer) finds an issue. If 'hey I'm Joe sixpack' calls and says "I think there's a mistake" it doesn't usually get pulled. So since the LHC stuff is peer reviewed, joe sixpack reviewed, and reviewed and commented on by just about everyone AND it still hasn't been pulled ...
It must have passed the 'stats test'.
And Yes, peer reviewed ain't what it used to be (probably never was), but that's not the point.
The LHC paper has been 'published'. It has been peer reviewed up the butt. It has not been withdrawn. It obviously then falls into the 'other' 999/1000. Like slashdot is fond of saying: there is nothing to see here, move along.
We've found that most employee's 'you are on good terms with' answer questions even after they leave. :).
:). We might delete/disable your personal acct on some critical servers for the last day or two you work (to check for stuff running as you that should be some other role/user) but we haven't changed root/admin password and we haven't tampered with your 'personal' computer. So in my opinion we haven't really removed your privs (you still have 'root', can't get too many more privs). You can go to a console and do whatever you want or have anyone in building logon as 'joe user' then escalate priv via root password (su is your friend on unix/linux and the network\administrator windoze acct still exists on the network). So that all can occur while you are still coming to work, if essential. Usually it is not an issue. we frequently look at your account stuff (visual 'oh crap' inspection), kill your acct in the AM and send you home at lunch with a 'call you if we need you' type deal.
Basically, the remove 'privs to ensure everything keeps on ticking' you state is the same as 'remove privs and send home' to us. If we are on good terms and we sent you home early, we aren't shy about calling/emailing if we get desperate.
Again, the difference we see is: send you home early with pay and we get desperate, more likely to help. Remove privs and make you come sit here anyway THEN we get desperate, less likely to help.
Yup, there is a risk you'll tell us to 'go fish' but it hasn't happened yet (in the whole couple of times we've had to call).
Now if your issue is: remove privs and keep them coming in for a couple days and if something breaks, have them fix it. We are in agreement, sort of. Difference is we don't remove privs
Seems to work for us.
OK, I guess I'll de-cloak and jump in. First background, I'm a CTO with a small software company. By small I'm talking about 30 bodies total. I mention this only to provide some data on my experience/viewpoint (CTO, not developer) and my environment (dev shop, small company) and even with all that said, remember YMMV.
OK, our policy falls in to two categories/buckets:
1 - your privs are removed and you are sent home with pay for the notice period, goodbye don't come in to work.
2- you keep all your privs and you continue to work, thanks for staying during the notice period.
Nothing else makes sense to us. Removing your privs and having you come in just creates a distraction while you talk to other staff, not useful to us.
As to whether you fall into bucket 1 or 2 is the result of conversations among management. Any doubt that you will play nice - goto bucket 1 immediately. Any doubt that you are really needed to complete work - goto bucket 1 immediately. If you both can and will contribute to the project and we do not expect any issues with you working during your notice (poaching employees, causing trouble, etc.) then go to bucket 2.
We have had people we assigned to bucket 1 that were great employees and I'd like to keep. They were not really needed for the project and we sent them home as sort of a last 'paid vacation' from us. No ill will, I'd hire them again. We've also sent people home and taken a hit on the project as the distraction, productivity, or trust factor outweighed the usefulness factor.
Removing your privileges and still having you come in makes absolutely no sense to me. Seems to be the worst of both worlds, you can't really be productive and the low work load can cause you to create distractions for other staff. I just do not get why they want to do that.
Hope that helps.
So how is knowing that he was talking about an application on DOS 2.0 in 1989 and that the question arose in 1990 going to help now. Or differently, assume it is on windows or linux now and it is in ASP or PHP; now provide information using that that helps him make a similar decision in 2030 when neither Windows nor Linux exists any more, hell the Internet as we know it will not exist. The point is "How do I make the general decision"/"What metric do I need to measure/compare in future decisions", NOT what is the recommendation for today. In future decisions he may not (nay, will not) be using the same environment as today.
Asking about the specific environment and providing specific information related to that environment doesn't really answer his question. It does make "our" job easier as we can all focus on the smaller specific issue and not have to deal with the general complex case. The "question" is what is the general case metric to determine obsolescent platforms. Or a better one would be a methodology that provides a numeric value rating the TCO of an architecture/platform. The question, of course, has no one answer and a general solution is not readily available.
What he will find is, the more general/vague the question, the more general/vague the answer has to be. OR the more the answer resolves only a discrete instance. We can all guess at a platform and provide data. If by "Platform" we mean web based multi-user application or client server application and we restrict our answers to that, we will get some form of response for that discrete 'instance'. We can further restircit to a language, ASP/PHP and provide a more specific answer. That answer will be useful until 20xx when that platform/language 'dies'.
We can try to answer the general question but the answer gets pretty complex/vague/academic. E.g., For any system/platform/entity we can decide a set of properties that are important, say flexibility, availability of resources (developers), etc. and assign weights (e.g., web sites that are relatively static need little implementation flexibility and have low resource consumption due to low change rates). We can add other variables we think important (e.g., lost opportunity costs) and add weights for those. All this continues until we have the massive untested model. The equation of which would contain numerous weighting factors and if written on paper would probably be massive enough to kill a cat (or not).
Now we can continue to run down this rabbit hole but I expect that little of practical value will emerge the other end. The key to the data the submitter wants is, as usual, in asking the correct question. General open ended questions are not known for generating highly useful, reusable, succinct answers. Highly specific questions are more apt to generate highly useful and possibly succinct answers but they frequently leave much to be desired in reusability. The best we can do is state this obvious issue to the original poster and suggest they rethink the issue and pose a question that strikes a correct balance for the information the submitter needs. I personally intend to conserve what few brain cells I have left and move on to other topics.
No he it NOT asking whether he should drop his platform. He is asking the more general question, how do you know when ANY platform has reached it 'drop' time'. You want to feed him a fish, he is asking to be taught to fish, big difference.
That breaks other stuff. The concept of +infinity -infinity is very useful. Otherwise series like x+2x+3x+4x+... converge to the same limit/value as x-2x-3x-4x-... as x->infinity. Kinda like saying all infinities are created equal which breaks lots of advanced math that relies on cantor set issues.
I'm all for simplifications but too simplified also causes issues. A long time ago my son was learning about sqrt. He asked the teacher about sqrt of negative numbers. The reply was "you can't take the sqrt of a negative". First, being a bright child, he ask me when he got home. His statement was, the teacher said you can't do sqrt of negative numbers, that doesn't seem right, math is usually more consistent than that, is it true? A short discussion of complex and imaginary numbers followed. Later in the year the teacher taught imaginary numbers.
The net result of this education was: my son learned to never assume the teacher was telling the whole truth (his statement). That whenever the teacher said something couldn't be done or was wrong it MIGHT just be wrong until the teacher decided to talk about it later. The teacher's attempt at simplification (in the short term) led to a lesson in the fact the teacher oversimplifies to avoid discussion and could not be trusted to tell the 'whole truth'. Not EXACTLY the lesson the teacher hoped for by oversimplifing (I assume).
So while simplification is good, oversimplification to make a point is not always good (and sometimes makes a different point).
I admit to not having read EVERY comment in this thread, so it is possible this is a dup but ...
... 1/-.5 ... 1/-.1 ... Is an stream of negative numbers increasing in absolute value e.g., -1, ... -2, ... -10, ... So as you take the limit of division buy zero from the negative side it goes to negative infinity. Now try:
... 1/.5 ... 1/.1 ... and get a stream of positive numbers approaching positive infinity.
:).
The guy is wrong. The reason divide by zero is undefined is simple:
1/-1
1/1
Try a graph.
Now explain how a single number/value/nullity can represent BOTH positive and negative infinity. Maybe if I create a function that approachs positive 5 from one direction and -5 from the other direction I can coin the 'fivish' as the solution and get my 15 minutes of fame as well.
Ah well, America seems to own crappy science these days I guess the Brits are entitled to crappy math
I think the issue is:
Think in-city clustered 12 story apt buildings. A simple building, 1 apt per floor, say a 1000 sq ft apt. So we have 12 'stacked houses' that have a single 'shared' 1000 sq ft roof. So 1000 sq ft of roof needs to support 12 'houses' of energy. Unlikely. THAT'S why the grid comes in. In clustered/dense city dwelling environments, roof top energy generation will be insufficient to cover the usage. We'll always need a 'grid' in that environment. How we fuel/supply the grid may be debatable but the existance is required due to the space/environmental constraints.
Those are copyright MS. Try freebasic www.freebasic.net for the 'free' version.
Oh, I agree that the smaller the blocks the easier it is to 'automagically' prove the code. I'll also agree that 20+ years of dev experience have taught me that small blocks are better even if we do not formally prove anything. Small blocks are more likely to be 'visually verified' by the developer and the follow-on developers, reducing bugs. If you can't hold the whole thing in your head, it is more likely to have issues. Compartmentalization is GOOD.
... If you can get a 30% reduction in errors per 1000 lines via a method but you need 2 times the lines you net MORE errors. Assume a statistical 10 errors per 1000 lines and a project of 10,000 lines, that's a likely 100 errors. Now the improved methodology statistically has only 7 errors per 1000 lines (30% better) but the project requires 20,000 lines yielding 140 errors. Now 10,000 of those lines MIGHT be provably correct due to structure, but if you DON'T prove them you have to count them (I agree if you truly prove them you can remove them from the estimate, yielding 70 likely errors).
But more code is more code, if you do not 'prove' the implementation. If compartmentalization reduces the errors per 1000 lines of code from X to Y, say 40% of the original value, you still can not claim a 60% reduction in bugs as the lines of code numbers are different in the different implementations.
To pull numbers out of the air
As I said: In the real world people aren't proving the code, they are claiming it COULD be proven. BZZZZT, thanks for playing. You now have to count the extra code at the lower statistical error rate AND you have to pay the performance penalty. Sometimes the lower rate and the additional code does NOT yield a lower predicted error count so why do it. Sometimes it does yield a lower predicted error count, at a performance hit, yielding a real world trade-off.
ALSO the compartmentalization frequently yields a 'lower exposure' per bug. In a mono kernel exposure is likely all of ring 0. In a microkernel an exposure is not always ring 0, sometimes ring 1 or 2, yielding a POTENTIALLY lower damage/risk. Whether it is ACTUALLY a lower risk depends on how the box/OS is used.
Basically, the whole threat model changes. This leaves you with an apples and pears comparison, not apples to apples. The trade-off needs to understood and an intelligent decision needs to be made. Consequently stating that a general purpose computing platform delivered with a microkernel that has lower performance is better than a monolithic kernel that has better performance is a useless statement (apples to pears).
If you understand the trade-off and the application environment you are applying it to, you can decide if you prefer the shiny apple or the shiny pear. It is NOT a choice between the shiny apple and the wormy apple.
If you want a shiny apple vs. a wormy apple, give me a microkernel environment with better performance than the mono kernel that has less lines of code and a lower statistical error count than the mono kernel and I'll be happy to make the obvious/easy choice.
Actually, I am not interested in proving the kernal correct, no one is in the real world (imo). I am interested in being assured that my PROCESSING is correct. While the micro kernel may be smaller (not equal to simpler) I still need to then look at all the services that the mono kernel supplied that are external to the micro kernel.
...) is a completely different animal. No one size fits all (shocking I know).
In all cases where a general computing OS is used the services list is probably the same (custom kernels are a different matter and a micro may strip down better, or not).
Yes you get SOME advantage in security as MAYBE less things run at ring 0 but still lots of stuff will run at elevated priviledge (ring 0 or 1 not 3) and you have more complexity.
In my opinion performance is one thing you care about, maybe less than security but you care. If the end-to-end security issues are even roughly the same between mono and micro kernels then performance trade-offs start to weigh in more (if I can have 1% better security at a 50% performance loss, do I want the trade? Where is the swap point, 5% security and 25% performance, 10 to 10, 50% better security at a 1% performance hit?). Oh, and yes the brighter reader is correct, it depends on my usage, does is not?
Micro kernel OSs are more complex end-to-end. More complex typically implies more code, more code typically implies more potential bugs. However, microkernels are more compartmentalized, better compartmentalization improves security (broken part A does not imply access to part B). So the intended usage of the kernel impacts its overall end-to-end security 'rating' for the application in comparison to a mono kernel.
A system that is only used by a single user, no network access, physical access restricted, user always runs a root, is probably as secure in either case (and a damn rare bird). A system that is naked on the net running 50 different publicly accessable services (web, DNS, mail, IM,
Life is full of trade-offs, this is just another one.
This is the age old argument, 'give a man a fish' or 'teach a man to fish'. Yes we can build roads/power plants/drugs for them, OR we can improve education so they can eventually build their own 'roads'. You appear to be of the 'give a man a fish' school while the media lab is of the 'teach a man to fish' school. You are both right and you are both wrong.
Solving the short term at the expense of the long term, is usually considered bad public policy (but usual public policy). Concentrating on the long term while people suffer is considered bad humanitarian policy (but watching people suffer is usual individual policy). I doubt we'll solve this debate here.
Ahhh, did anyone check the math. It looks a bit off. First he uses a 1.15 multiplier to account for 'other costs' THEN adds it to the loan value (i.e., interest oriented). If you read the endnote that is based on the fact that loans are for 115% of the value (payoff on old car?). How is that a legit 'cost' of the new hybrid car?
Second he is using the full cost of the hybrid. He is assuming that you dump a perfectly good car and buy a hybrid, NOT that you are bright enough to buy a hybrid when it is time to buy something. That is, he is assuming it is the full cost, not the incremental cost of the hybrid. While that MAY be a correct financial analysis, it is unlikely to be a real world analysis (IMO).
If I want a $22K hybrid and my other choice is a $18K car/SUV at 25MPG, then the 'additional capital expense' is $4K NOT $22K. $4K * 1.15 (assuming I use his magic math) is $4.6K incremental cost at 5.25% over 60 months that's about $88/mo in payment. Given the gas savings and higher trade in allowance, the case for a hybrid may be closer than he paints. Of course that assumes the competition for your car dollar is an SUV at 25 MPG if it is a small car at $15K and 30MPG then the hybrid case is less good.
The real issue is during a "I'm going to buy a new car, what will it be" purchase period. It is fair to deal with incremental costs and incremental improvements in gas mileage/trade-in value. As I read it, the article assumes a 'forced trade' at full cost, not incremental costs. I'm not sure that is a fair comparison.
Think easily distributable E-text books and CBT programs. In some areas a teacher comes way after a doctor or field irrigation (not saying it should, but it does). A cheap laptop with hand crank power that is full of e-books on first aid, farming, and basic 3-Rs education is useful as a community resource. (Think library and basic training) It is a one time cost not a recurring cost (OK, recurring every N years if you take care of it). A teacher with a roof, school, and materials is a recurring cost to the community.
Is the approach practical? I do not know. Probably depends a lot on the PC survival rate and quality of material available. I'm sure there are parts of this planet where a PC with a 5 year life at $100 ($20/yr) is a viable alternative to a paper based reference library and the one day a month a teacher might wander through.
If not, it is viable in the US where the average high school student in my area totes 20+ pounds of books between classes. Move to e-books and this low price reader and save the storage, transportation, distribution, and student agony of all that paper hauling.
Potentially 4 times the size folded up neatly and in my pocket so my old 'boomer' eyes can see the type. Think a foldable e-book that I can load with a couple of novels, carry in my pocket, and read on the train.
Add a tiny bit more cpu, an audio out jack, and an MP3 player. Now you have an e-book with MP3 player for the commuter.
Or take a PocketPc/Palm/PDA and add a mono video out jack for one of these and you have a decent e-book screen for a PDA.
Or use the PDA size device, add this screen, add Linux, add a wired and wireless ethernet device, add snort/tcpdump/... and you have a large screen (unfolded) network diag device with PDA keyboard (add USB jack for USB keyboard possibly available at destination).
Bottom line, I prefer a backlit color LCD too. I also prefer long battery life (days not hours), and larger type for my older eyes. Life is full of trade offs. I'll give up the color screen for more battery life and larger print in a commuter device or network diag device.
It is not a laptop and will not replace a laptop in today's world until at least 8-16 bit color is available. Stop thinking laptop and start thinking large screen low weight appliance with longer battery life.
(Note: think a mono laptop (OK 2 bit greyscale) that had old mono PDA battery life (days/weeks), weighed less than 8 ounces, was PDA size with 'fold out' keyboard and wireless would probably have some market share as an e-book/light net surfing (text)/commuter appliance.)
(Note: I like Linux, I run Linux, the company runs Linux in the infrastructure. The following is a business discussion NOT a personal preference.)
OK, go to Dell chose a Linux PC and note the price. Now go to a windows XP Home based PC and 'upgrade' the config to match the Linux PC from a speed/memory/disk/video perspective. Last time I did this the price difference was ~$40 on an $800 PC (5%). Remember that while XP is $200 retail it is ~$30-50 OEM (hey $40 times a million PC per year per 'Dell' adds up).
On Walmart.com I see a celeron D 2.8Ghz 256meg 80G drive Win XP Pro for $407. A celeron D 2.8 Ghz, 128 Meg, 40G drive, no OS is $78 cheaper (note smaller drive and less ram on No OS and the Win box is Win PRO not home). Walmart does not really do support (manufacturer support) so it is less likely to mark up for support costs (the manufacturer to walmart costs MAY already have a markup for it).
So the 'manufacturer' has double the OS test/certification costs to reduce his retail price ~$50 or ~5%. For this effort he gets entry into a small market space. THAT IS THE MAJOR ISSUE. If I have to test on multiple OSes and support multiple OSes I need a reasonable return on that cost. Right now it is not there.
The cost of training the support staff to support Linux doubles or triples my support training costs. I now have N sets of drivers to track and make available on my support web site. I have extra sales training (I want to use Linux and I want to do X, what PC model and software should I buy).
The expectation as has been stated here a zillion times is the PC will 'cost less'. So for my ~$50 in price difference I MIGHT sell a few PCs to the currently small Linux market place. For that I have to control my support costs and sales costs to market two different OS systems in order to keep a profit. (note: assumption, the $50 difference in price is roughly what the OEM cost of Windows is, does anyone know the OEM pricing for large retailers like Dell?).
That's on an entire PC. For a network card manufacturer selling product at ~$20 retail his 'profit' is probably a couple bucks per card, why should he assume the extra cost burden of doing driver dev and testing twice and supporting them?
I can 'almost' double my training and software tracking/'driver dev' costs and support the whole market or get ~90% of the market at half that cost.
Please explain again why the business guys are not making a good business decision?