I'm not all that far from that assembler programmer. First, I've done that: programmed assember on several microprocessors, and on DEC 10s and 20s. I've an undergraduate computer engineering degree and an MS in computer science.
But MSFT products "scare" me. Oh, I'm not quite so bad as she. But I know that an error or problem on an MSFT environment is provided in "user friendly" form, with details and solutions hidden. How one problem is solved may be completely unrelated to how another problem is solved, regardless of how similar the problems might appear to be.
This is in contrast to my typical experience with other environments. In those environments, there's a level of consistency. A little knowledge and some guesswork normally provides a solution. I think of these as "predictable". MSFT products are far from predictable.
I'm not so much scared of breaking something as I am scared of being sucked into something which would require hours or days to learn enough to fix, and but where that acquired knowledge would be useless in any other situation. That is, it would suck up my time with extremely limited reward.
I might attribute this to a lack of training on my part. After all, I've only done things like write device drivers in UNIX, or write interrupt-based software which ran "under" DOS environments on an 8088 CPU. I've taught programming in various languages, including Java, Perl, C, C++, SQL. I've taught OO and DB design concepts, sound programming practices, and some GUI programming. In my spare time, I fly small Cessnas. So I've a lot to learn yet.
But people with certification in these MSFT environments seem, if anything, more puzzled and more inefficient in their use of time solving problems.
In reading the original message (included at the bottom of the later message), I think that this has nothing to do with inbound spam. Instead, I believe that AT&T is about to block its clients from accessing port 25 on servers other than those in a defined list.
This doesn't address the problem of AT&T users receiving spam (except indirectly). Instead, it is addressing the problem of AT&T users sending spam. More likely, this is addressing the problem of poorly configured and virus-infected machines belonging to AT&T clients being used as relays of spam.
This is likely in response to the "stealth spamming" that's becoming more popular: hijacking machines via virus for use as SMTP relay, DNS server, and web server. [For those interested, there's been a fair bit of NANOG discussion of this recently under the subject of "Wired mag article on spammers playing traceroute games with trojanedboxes".]
Yes, and if you want to stop rape, women shouldn't grow tits, take baths, wear make-up, dance, etc... Back to burkas for them cause God forbid they tempt men.
An overly extreme example, but that's fine.
I see it very differently. The "don't use MSFT" analog to stopping rape involves learning karate, carrying mace, having a decent legal system, properly funding police services, etc. In other words: proactive protection. Of course, educating people that it's a Bad Thing applies to both cases.
Your example would be more along the lines of "disconnect from the Internet".
With respect to your applications, have you spoken with these vendors? Have you suggested that they make their products available for decent platforms? Perhaps they've received this request from just one too few people for them to make the leap, and you'd be the one to cause the trigger to be pulled.
It's still an interesting point. Given the number of different systems - even if there's duplication of tasks accomplished - one would expect a much higher exploit rate. Normally, this argument is made as to why MSFT products - with all their bundling - have so many problems. But while Linux itself is cleaner than MSFT environments, the distributions ship a lot more than just the OS.
And what's wrong with that duplication anyway? We often call that "choice". This may be a foreign concept in many cultures and computing platforms, but some of us relish such arcane concepts.
You might not know that the maintenance is being performed. For example, if your NOC uses HSRP (or some equivilent) to allow routers to back up one another, the NOC could take one router at a time out of service, upgrade it, and put it back, all w/o interrupting service.
Your NOC does have something in place which supports (1) pulling routers out of service for maintenance and (2) routers that fail on occasion, right?
One of the critical things to being happy is being able to make a living doing something that you love to do, that expresses you to yourself and to others, and that perhaps they don't even have to pay you to do.
One can express into anything. Have you eaten at a restaurant? Then you know the difference between a waitperson that is being the best waitperson he/she can be, and a waitperson that's just waiting to be somewhere else. This can make the difference between a good or bad meal.
Most of the people that drive the buses from my town into NYC are at best quiet and too often downright surly. But one or two are so nice and so friendly that articles are written in the local newspaper about them. That's the type of difference they make to commuters' mornings and evenings.
In fact, I'll correct what I wrote above. One cannot help but express into anything one's doing. The question is: will you express excellence and satisfaction, or will you express the desire to be doing something else?
It's too bad there isn't an HTML tag for "emphatic agreement"; I'd be using it here.
I came to this philosophy via a different route. I attend their workshops in NYC, but you can find their book here. It does share a lot with what little I know of Buddist thinking, but comes at satisfaction from a different direction.
The pursuit of excellence can become at habit. More, I've found that this is where one finds satisfaction. It's this pursuit, rather than the endeavor in which the pursuit is practiced, that yields satisfaction.
> Wow, did you ever miss out on a free marketing opportunity.
What would you expect? This fellow cannot even figure out how to run a business providing "unique content" to interested users.
Why does it need to be free? If your content is really both unique and useful, it's valuable. Charge for it. If people pay, you're right. If not, you weren't going to make it anyway.
The model for trying to steal attention is doomed to failure. VCRs and their descendents are impacting commercial viewing on television, and ad-blocking is impacting the Internet equivilent. So stop whining and find a new model.
I'm actually quite eager to see what will occur with TV. I'm wondering about product placement on shows, or perhaps going back to real sponsorship that was used in TV's early days.
What you're seeing is the distinction between authentication and authorization. Unfortunately, certain parties do not see the difference. Of course, these are the same parties that give us systems where the default behavior (and, in many cases, the only possible behavior) is to authorize *any* action given *any* authentication.
In fact, I think this a bit of a smoke screen. The law the DMA will really push is one which (1) prohibits an ISP from blocking "nonspam", (2) defines spam unreasonably (ie. requires forged headers, or some such), and (3) prevents ISPs from kicking off ISP clients that put nonforged headers on their "nonspam" spam.
I don't think the trigger of this change is the ever-increasing crap filling our mailboxes. Instead, tools like SPAMHAUS and SPEWS are too damned effective for the DMA's liking. Removing ISPs' ability to protect their clients and the rest of the Internet is the ultimate DMA goal.
What I've not seen (yet) in this discussion are the issues of "monopoly" and "public good".
Part of a government's role is to protect the functioning of the marketplace (however that is defined in that given region of governance). To much of the western world, this includes preventing abuses by monopolies.
In the US, we've the DOJ which takes on potential abuses in court. But that's just one possible approach. Another approach is a more market-driven one. That second approach makes a lot of sense to me, as it doesn't get a second system (the legal system) involved in the first (the economic system).
This second approach involves helping to cause the investment necessary to break (or neuter) the monopoly. Recall that a monopoly is rarely based upon absolute prevention of competition. It is more typically merely a matter of economic barriers which require too much investment to compete.
No company may be able to afford that investment. But a government, less tied to short term cycles and immediate returns on investment, may.
More, governments also serve as a conduit through which long term projects for the public good may be funded. It may not make sense for a company to invest in the building of a dam, but it may make sense for a government to do so.
I see the breaking of monopolies as a public good akin to building dams, Internets, and airports, protecting habitates, and the like.
Given the nature of the investment, and given the public good intended by the result, I see this as a very natural role for a government to play.
Perhaps the government can undo the corporate aggregation that may very well have been funded by funny accounting. The Internet could use a little more competition and smaller backbones with more reasonable peering requirements.
There's been a lot of traffic here (and elsewhere) on "how could this happen" and "what greedy SOBs they are." But I've a concern/issue/idea about which I see nobody talking.
Have I merely missed it?
Worldcom's been fudging numbers for how long? Worldcom has been buying Internet players for how long? Worldcom has been increasing peering requirements for how long?
From what I can see, Worldcom's creative accounting may have helped shape today's connectivity market. And where I write "shape", I mean "shrink". I don't refer to a reduction of bandwidth, but instead of competition.
And as Worldcom increased in market control, they made peering increasingly difficult/expensive. So routes, as others have mentioned here, have become interesting.
All of which may have come from Worldcom's cooking of the numbers.
So is there any possibility of a restored market? That is, might Worldcom be forced to disaggregate into the separate backbones, phone companies, network providers, etc. out of which it was formed? Or, at least, might the more recent purchases be returned to an independent life?
Even if your redundant connection is with some completely independent provider, do you know who owns - and who is running - the fiber over which that redundant link travels? What about the fiber that your independent provider uses to connect to its upstreams/peers?
Let's take this in a slightly different direction. I don't have dialup. I have several T-1s and a wireless link on my roof. I'm exchanging routes via BGP over all these links, running my own ASN.
And I still deal with the "monkeys" on one backbone. Do you know how difficult it is to get a message to an engineer with details regarding an error in his router's configuration if it passes through a person that cannot spell "BGP"?
Happily, this isn't always true. Two of my connections are to companies where I can speak to Engineers. Problems are resolved in minutes (aside from the rare situation where the phone company is involved, but let's not go there {8^).
But I had an outage on a pair of T-1s for *5* days because the people moving my circuits never bothered to move my BGP sessions to the new router on their end.
ISPs - even the large backbone companies - are trying to get away with a level of support that is offensive. What we need to do is vote with our dollars. I'm moving away, for example, from that company that gifted me with a 5 day outage and numerous conversations with illiterates. What are you doing?
Embedding type information within variable names is an artifact. It stems from poor coding practices, which permit great "distance" between declaration and use. It stems from poor tools, where these tools don't facilitate locating a declaration.
This practice is also inefficient, potentially very so, in that redundancy is being increased. Worse case, some function's name must be changed because of a change in the return type (or argument list).
There are such changes - depending upon the language being used - that will not require alterations in the calling code. For example, if a return of type T is changed to return a type derived from T (which therefore supports the full T interface), then callers remain uneffected. But if the return of type T was embedded within the function name, then all calling code must be changed.
As true as this is, I find comments describing "why not" even more useful. If I know what was avoided - or how/why some other approach failed - I am saved the cost of repeating that effort.
Apparently, it is not as easy as one would expect.
I watched someone corrupt a machine by installing some patch from MS. It would boot straight into the BSOD. He had to do manual tweeking from the CDROM to get it to work (a less savvy admin would no doubt have just reinstalled).
With a UNIX, I can have multiple kernels - and even root partitions. I can make my choice at runtime. Do MS environments have this feature?
This is a good point. You need to see - and explain - this from a business perspective. As someone else mentioned, it sounds like what you're doing is subject to automation. But when you propose that, do so not for the purpose of "prettier code", but more rapid production.
In other words, you'd be able to handle *more* clients for the same cost. In fact, you'll not only be able to handle more of the cookie-cutter stuff (which will, btw, improve in return for your form once this is automated), but you'll also be leaving yourself (and other developers) free for the non-cookie-cutter work.
It would also help if you've post-development cost factors to which you can point. There are many different forms these can take. How much does it cost to maintain the systems your clients are using, how much does it cost to add new features, how likely are clients to shift asway from your products - and therefore your company - as they grow or change, etc.
One trick I've often used is simply a special case of iterative improvement. Each time we reuse a cookie cutter, we learn more about it. We therefore improve it. Perhaps we add a feature, or make it easier to extend in a new direction, or improve the automation of cutting the cookies. The cost of this work is covered by the client, but all clients - past and future - receive the benefit of the work. That translates to your company receiving the benefit of the work.
This can actually be quite challenging. You're not just cutting cookies, but improving all of the tools used. This then takes you in the direction of better automation, etc., that I discussed above.
There's really a lot of room here for you to maneuver. But just remember to see and explain things from a business perspective. Unless you've technically savvy managers that will perform the tech->business translation themselves, you need to explain things to them in a language they understand.
I'm sorry, but I've no idea about what specific firm you're commenting. Mine was more of a general issue than anything specific to a given company. I certainly don't disagree that some firms have policies which, likely unintentionally, cause people to move as they gain experience.
This is *very* old news to me. I recall my father being in such a position over 20+ years ago. His managers were displeased with the situation, but lacked the ability to do anything. So rather than manage people making more than he was, he moved on.
I'd a similar situation at a large telecom company. I was a consultant to them, but willing to become an employee. All they needed to do was give me the title that I wanted, and I was willing to take a significant salary cut (I was young {8^). But the company policy prevented this, so they instead kept me on at the higher rate.
But I still think that there are too many cases where people change jobs just to change, because - at least recently - that's looked better on the resume. Hopefully, that's changing.
But there's a cost to hiring entry level, and that cost has become "worse" lately (although perhaps, post-bubble, this is less true). The issue is whether the soon-to-be-experienced person will stay or leave. At least over the past few years, it has been considered chic to bounce from job to job. That makes investing in training a little hard on the employer.
Note, however, that the ISP did not just take down, or block access to, the offending material. What they did was deny services to their client. This would have no effect on the material supposedly posted.
The proper action would have been to issue a USENET cancel of the offending article.
Slashdot Mirror
I'm not all that far from that assembler programmer. First, I've done that: programmed assember on several microprocessors, and on DEC 10s and 20s. I've an undergraduate computer engineering degree and an MS in computer science.
But MSFT products "scare" me. Oh, I'm not quite so bad as she. But I know that an error or problem on an MSFT environment is provided in "user friendly" form, with details and solutions hidden. How one problem is solved may be completely unrelated to how another problem is solved, regardless of how similar the problems might appear to be.
This is in contrast to my typical experience with other environments. In those environments, there's a level of consistency. A little knowledge and some guesswork normally provides a solution. I think of these as "predictable". MSFT products are far from predictable.
I'm not so much scared of breaking something as I am scared of being sucked into something which would require hours or days to learn enough to fix, and but where that acquired knowledge would be useless in any other situation. That is, it would suck up my time with extremely limited reward.
I might attribute this to a lack of training on my part. After all, I've only done things like write device drivers in UNIX, or write interrupt-based software which ran "under" DOS environments on an 8088 CPU. I've taught programming in various languages, including Java, Perl, C, C++, SQL. I've taught OO and DB design concepts, sound programming practices, and some GUI programming. In my spare time, I fly small Cessnas. So I've a lot to learn yet.
But people with certification in these MSFT environments seem, if anything, more puzzled and more inefficient in their use of time solving problems.
In reading the original message (included at the bottom of the later message), I think that this has nothing to do with inbound spam. Instead, I believe that AT&T is about to block its clients from accessing port 25 on servers other than those in a defined list.
This doesn't address the problem of AT&T users receiving spam (except indirectly). Instead, it is addressing the problem of AT&T users sending spam. More likely, this is addressing the problem of poorly configured and virus-infected machines belonging to AT&T clients being used as relays of spam.
This is likely in response to the "stealth spamming" that's becoming more popular: hijacking machines via virus for use as SMTP relay, DNS server, and web server. [For those interested, there's been a fair bit of NANOG discussion of this recently under the subject of "Wired mag article on spammers playing traceroute games with trojanedboxes".]
Any citations for this? I'd love to be able to show this to MSFT fanatics.
Yes, and if you want to stop rape, women shouldn't grow tits, take baths, wear make-up, dance, etc
An overly extreme example, but that's fine.
I see it very differently. The "don't use MSFT" analog to stopping rape involves learning karate, carrying mace, having a decent legal system, properly funding police services, etc. In other words: proactive protection. Of course, educating people that it's a Bad Thing applies to both cases.
Your example would be more along the lines of "disconnect from the Internet".
With respect to your applications, have you spoken with these vendors? Have you suggested that they make their products available for decent platforms? Perhaps they've received this request from just one too few people for them to make the leap, and you'd be the one to cause the trigger to be pulled.
It's still an interesting point. Given the number of different systems - even if there's duplication of tasks accomplished - one would expect a much higher exploit rate. Normally, this argument is made as to why MSFT products - with all their bundling - have so many problems. But while Linux itself is cleaner than MSFT environments, the distributions ship a lot more than just the OS.
And what's wrong with that duplication anyway? We often call that "choice". This may be a foreign concept in many cultures and computing platforms, but some of us relish such arcane concepts.
You might not know that the maintenance is being performed. For example, if your NOC uses HSRP (or some equivilent) to allow routers to back up one another, the NOC could take one router at a time out of service, upgrade it, and put it back, all w/o interrupting service.
Your NOC does have something in place which supports (1) pulling routers out of service for maintenance and (2) routers that fail on occasion, right?
What prevents this same action from occurring during the discovery phase?
One of the critical things to being happy is being able to make a living doing something that you love to do, that expresses you to yourself and to others, and that perhaps they don't even have to pay you to do.
One can express into anything. Have you eaten at a restaurant? Then you know the difference between a waitperson that is being the best waitperson he/she can be, and a waitperson that's just waiting to be somewhere else. This can make the difference between a good or bad meal.
Most of the people that drive the buses from my town into NYC are at best quiet and too often downright surly. But one or two are so nice and so friendly that articles are written in the local newspaper about them. That's the type of difference they make to commuters' mornings and evenings.
In fact, I'll correct what I wrote above. One cannot help but express into anything one's doing. The question is: will you express excellence and satisfaction, or will you express the desire to be doing something else?
It's too bad there isn't an HTML tag for "emphatic agreement"; I'd be using it here.
I came to this philosophy via a different route. I attend their workshops in NYC, but you can find their book here. It does share a lot with what little I know of Buddist thinking, but comes at satisfaction from a different direction.
The pursuit of excellence can become at habit. More, I've found that this is where one finds satisfaction. It's this pursuit, rather than the endeavor in which the pursuit is practiced, that yields satisfaction.
> Wow, did you ever miss out on a free marketing opportunity.
What would you expect? This fellow cannot even figure out how to run a business providing "unique content" to interested users.
Why does it need to be free? If your content is really both unique and useful, it's valuable. Charge for it. If people pay, you're right. If not, you weren't going to make it anyway.
The model for trying to steal attention is doomed to failure. VCRs and their descendents are impacting commercial viewing on television, and ad-blocking is impacting the Internet equivilent. So stop whining and find a new model.
I'm actually quite eager to see what will occur with TV. I'm wondering about product placement on shows, or perhaps going back to real sponsorship that was used in TV's early days.
What you're seeing is the distinction between authentication and authorization. Unfortunately, certain parties do not see the difference. Of course, these are the same parties that give us systems where the default behavior (and, in many cases, the only possible behavior) is to authorize *any* action given *any* authentication.
It's a resource available for theft. Liquor stores are a part of armed robbers' models too.
No kidding.
In fact, I think this a bit of a smoke screen. The law the DMA will really push is one which (1) prohibits an ISP from blocking "nonspam", (2) defines spam unreasonably (ie. requires forged headers, or some such), and (3) prevents ISPs from kicking off ISP clients that put nonforged headers on their "nonspam" spam.
I don't think the trigger of this change is the ever-increasing crap filling our mailboxes. Instead, tools like SPAMHAUS and SPEWS are too damned effective for the DMA's liking. Removing ISPs' ability to protect their clients and the rest of the Internet is the ultimate DMA goal.
What I've not seen (yet) in this discussion are the issues of "monopoly" and "public good".
Part of a government's role is to protect the functioning of the marketplace (however that is defined in that given region of governance). To much of the western world, this includes preventing abuses by monopolies.
In the US, we've the DOJ which takes on potential abuses in court. But that's just one possible approach. Another approach is a more market-driven one. That second approach makes a lot of sense to me, as it doesn't get a second system (the legal system) involved in the first (the economic system).
This second approach involves helping to cause the investment necessary to break (or neuter) the monopoly. Recall that a monopoly is rarely based upon absolute prevention of competition. It is more typically merely a matter of economic barriers which require too much investment to compete.
No company may be able to afford that investment. But a government, less tied to short term cycles and immediate returns on investment, may.
More, governments also serve as a conduit through which long term projects for the public good may be funded. It may not make sense for a company to invest in the building of a dam, but it may make sense for a government to do so.
I see the breaking of monopolies as a public good akin to building dams, Internets, and airports, protecting habitates, and the like.
Given the nature of the investment, and given the public good intended by the result, I see this as a very natural role for a government to play.
Perhaps the government can undo the corporate aggregation that may very well have been funded by funny accounting. The Internet could use a little more competition and smaller backbones with more reasonable peering requirements.
There's been a lot of traffic here (and elsewhere) on "how could this happen" and "what greedy SOBs they are." But I've a concern/issue/idea about which I see nobody talking.
Have I merely missed it?
Worldcom's been fudging numbers for how long? Worldcom has been buying Internet players for how long? Worldcom has been increasing peering requirements for how long?
From what I can see, Worldcom's creative accounting may have helped shape today's connectivity market. And where I write "shape", I mean "shrink". I don't refer to a reduction of bandwidth, but instead of competition.
And as Worldcom increased in market control, they made peering increasingly difficult/expensive. So routes, as others have mentioned here, have become interesting.
All of which may have come from Worldcom's cooking of the numbers.
So is there any possibility of a restored market? That is, might Worldcom be forced to disaggregate into the separate backbones, phone companies, network providers, etc. out of which it was formed? Or, at least, might the more recent purchases be returned to an independent life?
Even if your redundant connection is with some completely independent provider, do you know who owns - and who is running - the fiber over which that redundant link travels? What about the fiber that your independent provider uses to connect to its upstreams/peers?
Let's take this in a slightly different direction. I don't have dialup. I have several T-1s and a wireless link on my roof. I'm exchanging routes via BGP over all these links, running my own ASN.
And I still deal with the "monkeys" on one backbone. Do you know how difficult it is to get a message to an engineer with details regarding an error in his router's configuration if it passes through a person that cannot spell "BGP"?
Happily, this isn't always true. Two of my connections are to companies where I can speak to Engineers. Problems are resolved in minutes (aside from the rare situation where the phone company is involved, but let's not go there {8^).
But I had an outage on a pair of T-1s for *5* days because the people moving my circuits never bothered to move my BGP sessions to the new router on their end.
ISPs - even the large backbone companies - are trying to get away with a level of support that is offensive. What we need to do is vote with our dollars. I'm moving away, for example, from that company that gifted me with a 5 day outage and numerous conversations with illiterates. What are you doing?
Embedding type information within variable names is an artifact. It stems from poor coding practices, which permit great "distance" between declaration and use. It stems from poor tools, where these tools don't facilitate locating a declaration.
This practice is also inefficient, potentially very so, in that redundancy is being increased. Worse case, some function's name must be changed because of a change in the return type (or argument list).
There are such changes - depending upon the language being used - that will not require alterations in the calling code. For example, if a return of type T is changed to return a type derived from T (which therefore supports the full T interface), then callers remain uneffected. But if the return of type T was embedded within the function name, then all calling code must be changed.
Expensive.
As true as this is, I find comments describing "why not" even more useful. If I know what was avoided - or how/why some other approach failed - I am saved the cost of repeating that effort.
Apparently, it is not as easy as one would expect.
I watched someone corrupt a machine by installing some patch from MS. It would boot straight into the BSOD. He had to do manual tweeking from the CDROM to get it to work (a less savvy admin would no doubt have just reinstalled).
With a UNIX, I can have multiple kernels - and even root partitions. I can make my choice at runtime. Do MS environments have this feature?
This is a good point. You need to see - and explain - this from a business perspective. As someone else mentioned, it sounds like what you're doing is subject to automation. But when you propose that, do so not for the purpose of "prettier code", but more rapid production.
In other words, you'd be able to handle *more* clients for the same cost. In fact, you'll not only be able to handle more of the cookie-cutter stuff (which will, btw, improve in return for your form once this is automated), but you'll also be leaving yourself (and other developers) free for the non-cookie-cutter work.
It would also help if you've post-development cost factors to which you can point. There are many different forms these can take. How much does it cost to maintain the systems your clients are using, how much does it cost to add new features, how likely are clients to shift asway from your products - and therefore your company - as they grow or change, etc.
One trick I've often used is simply a special case of iterative improvement. Each time we reuse a cookie cutter, we learn more about it. We therefore improve it. Perhaps we add a feature, or make it easier to extend in a new direction, or improve the automation of cutting the cookies. The cost of this work is covered by the client, but all clients - past and future - receive the benefit of the work. That translates to your company receiving the benefit of the work.
This can actually be quite challenging. You're not just cutting cookies, but improving all of the tools used. This then takes you in the direction of better automation, etc., that I discussed above.
There's really a lot of room here for you to maneuver. But just remember to see and explain things from a business perspective. Unless you've technically savvy managers that will perform the tech->business translation themselves, you need to explain things to them in a language they understand.
I'm sorry, but I've no idea about what specific firm you're commenting. Mine was more of a general issue than anything specific to a given company. I certainly don't disagree that some firms have policies which, likely unintentionally, cause people to move as they gain experience.
This is *very* old news to me. I recall my father being in such a position over 20+ years ago. His managers were displeased with the situation, but lacked the ability to do anything. So rather than manage people making more than he was, he moved on.
I'd a similar situation at a large telecom company. I was a consultant to them, but willing to become an employee. All they needed to do was give me the title that I wanted, and I was willing to take a significant salary cut (I was young {8^). But the company policy prevented this, so they instead kept me on at the higher rate.
But I still think that there are too many cases where people change jobs just to change, because - at least recently - that's looked better on the resume. Hopefully, that's changing.
But there's a cost to hiring entry level, and that cost has become "worse" lately (although perhaps, post-bubble, this is less true). The issue is whether the soon-to-be-experienced person will stay or leave. At least over the past few years, it has been considered chic to bounce from job to job. That makes investing in training a little hard on the employer.
Note, however, that the ISP did not just take down, or block access to, the offending material. What they did was deny services to their client. This would have no effect on the material supposedly posted.
The proper action would have been to issue a USENET cancel of the offending article.