I was quite surprise in doing some OCCAM programming on Transputers, by the level to which you could turn processes into something that ran well concurrently.
Part of the problem we have at the moment is that setting up threads, or forking, is a tedious business and prone to debugging hell. In OCCAM you could write your procedures and where and when you wanted them to run serially or in parallel, put a case statement like structure that was headed with SER or PAR.
Seeing this, it was easy to think of new ways in which you could decompose existing problems and then parallel-ize them to make the most of the hardware.
C++ has been a pile of steaming crap for better than 10 years. Review the success of Java in the period. Cross refernece with the continued us of C to do most (non-windows) systems programming.
As for the Author being the Chair of the standards committee, this just goes to prove we all make mistakes once in a while.
- The IQ difference between a worm and a human is such that it allows the worm to burrow deeply and get out of trouble.
I dunno! I think your Pentium 133 is running slower than most. If todays 'average' PC is a 3GHz screamer that would make your new machine 22 times faster than the old Pentium 133 paper weight.
There has been some improvement in how fast a single instruction operates on the latest Intel chip generations but nothing to get you being 'a few hundred' times quicker.
The Athlons are a different game, AMD did tweak the instruction set to run faster, hence their nominal speed ratings are very different from their actual speed ratings. The nominal rating is what an Intel chip would (surely) have to be clocked at to achieve the throughput. But still not to achieves hundreds of times the performance.
Personally, I need all the speed I can get. I need to get into the top 2 percentile of all Seti@home users... NOW!!! And that means Powwa!! Quad-10GHz processors... mummm!
I'm not sure our ICBM's are accurate enough for what you want. A Minuteman III had a CEP of about 300 ft. That mattered little if the 500Kt warhead went off. But landing 500lb's of dummy warhead concrete 300ft from a house is not going to do anything but make a loud noise and make a small hole.
But that makes a great project for these guys to get on - improving our ICBM accuracy so we can put concrete blocks in the nose instead of thermonuclear devices. Yes, another way to spend billions!
it is not physically possible to expect a reasonable chance
It is either possible to hit something or it is not. Throwing chance into the equation is fence sitting.
While the tests have proven that the task of vaporizing an incoming warhead is hard, they have also proven it to be possible
At a tactical level, making sure that the right bits are hit, and not decoys, becomes something that will eventually fall into place. However, that is possible - we do it in other areas of missile interceptions.
What I object to in all of this, is that we will not be attacked by a missile force, therefore this method of defence is useless, and an extreme waste of money. If China or Russia wanted to nuke us with missiles, this system would not shoot down enough. And if North Korea wanted to shoot at us, the return address is easily read. If a terrorist launched a scud missile from a ship in the Carribean, it couldn't be detected or intercepted anyway - it would be coming from the wrong direction and be too low
All people are different. Some like symbols, others like patterns, others smells. To say that people who can't remember passwords are stupid, is a sign of limited thinking on your part.
I barely remember peoples names, yet I have met people that have remembered the name of everyone they have ever met.
I once memorized the hex op codes for the 8086 instruction set and could look at a memory dump and generally work out what it did without reference to a manual, or the disassembler. But I can't save multiple passwords in my head for a nanosecond.
Worked at UPS in New Jersey back in 1993. One particularly ugly morning - there were many ugly days aqt UPS - the automatic password expiration process kicked in. It threw me off my session and demanded that I change my password. I myself was in an ugly mood, so selected as my password "fu*kyou". This seemed appropriate for a place like UPS, at that time. I logged back on and continued to do my work.
Within an hour I was kicked off again, and couldn't log back on. I called systems and they told me the account was disabled and I should talk to my manager. My Manager? I saw the manager and he called, they said they needed his authorization to re-enable my account...
I called systems again. This time they were a bit more willing to talk. It turns out the passwords were held by the UPS mainframe in clear text... and monitored. The sys admin girl seemed a little shocked on accessing my account. I asked, "Was it the password?" "Oh! Yes!" was the reply.
So, security be damned.
I have to say, I have never again used that password!
I concur. I tried both Solaris 7 and 8 as x86 and they were worse that anything I have ever tried - including OS/2 1.2. There is NO driver support for Solaris. It may be great as a server system with a few proprietary hard drive controllers, but forget doing any desktop stuff with it.
Sun's concept of desktop is also a little warped. I don't think they have ever got their heads around the fact that people unplug their PC's from the mains and wander around with them - still on.
I'm sorry but I think you are being the Nimrod here. With the downturn having taken up the better part of 4 years and.NET having only been a viable option for most people over the last 2 years, there is no way that there are that many people with both legacy language skills (C/C++) and new techniques (C#, VB.NET) in a production environment.
This is compounded by your request for product roll out. Few people anywhere have.NET product roll out experience.
You are trying to do more with less people, but the combination of skills that you are requesting to make up this deficiency just may be counter intuitive and perhaps even mutually exclusive.
Look at people with Java or C++ skills and say you will train in.NET if necessary.
Banging a stake in the ground and telling the candidates they must be HERE! Today! is a little presumptuous.
Maybe we could call this.... an appruntecip... no no... how about an... an "apprenticeship!" Yes! that sounds like a good word.
But they will need 3 years experience before they can start...
Re:Had to completely uninstall the Preview Release
on
Firefox 1.0 Released
·
· Score: 2, Informative
Yup, you have to completely uninstall the prior version. The uninstall WILL NOT remove your prior settings, they are retained. It's an act of faith, just Believe! Brothers and Sisters!
Ahem... Sorry for that outburst, I am a survivor of the vi / emacs religious wars and sometimes I get flash backs.....
Yup! This is probably the way to go. Many other solutions mentioned here will either not work or will dissolve the PCB or its lacquer coating, rendering the electronics useless.
Of course TCTFE is a carcinogen, is mostly banned in all civilized parts of the world and does a number on your brain if you sniff enough of it, but hey, it's a great cleaner.
ICL, that fine (Anglo-)Japanese computer manufacturer - and my first employer - had a whole mainframe they salvaged from a fire. It was taken apart completely, the whole thing then dumped in this cleaner, dried, and re-assembled. The machine ran better and more reliably that any ICL mainframe of the day, though that is not necessarily saying much.
I read the conclusion of this story and have a comment: These were not hobbist computers and parts. Instead they were really what at the time were called Business Systems. Byte in those days was subtitled something like "For Small and Business Systems" making a case for the classification of these computers and their use. I worked in 1981 for a small company selling "Business Systems" (ZR Computers). The $6000AU disk drive was for businesses, not the home computer owner. NASCOM-1's were for Home buyers, as were many smaller Z80 / CP/M systems (Trash 80 and Sharp MZ80's come to mind).
My first job for ZR was to work with a 'network' of 3 Commodore 8096 PET's that used a single Commodore 10MB disk unit. 2 computers used the the IEEE-488 bus to access the Hard drive, the other used the network to communicate with one of the drive connected PET's to get it's data. I called it Master/Slave. But what did I know about political correctness. This drive was a 5000 Pound UK drive unit, at a time when the Pound was worth something (!)
The software I wrote, a machine tool parts catalogue, was billed out at 400 pounds UK (say $800) and took me about 6 months to write. I was paid 3000 pounds / annum, therefore the software cost ZR, even at my stunningly low wage rate, 1100 pounds. Business idiocy is not something new!!
RPN... Somewhere around here is my old Sinclair Scientific (circa 1977). If only the on/off switch still worked. Clive Sinclairs early UK RPN calc. And it was a pretty thing, too, all in white.
But it gets me thinking. I have a roman numeral calculator on the website, could I do an RPN (Reverse Ptolemaic Notation... Poland didn't exist in 200BC) Roman Numeral Calculator? Hmmm! Now there's a project!
II II + IV
I'm all excited!!!
Better, may be not; warm and friendly, absolutely
on
Retro Vision
·
· Score: 1
I don't doubt that there are better programs now, and that some of the shows we watched in years of old were junk, but every so often it is just great to come across something that you loved and makes you feel warm and fuzzy once again. And while not of the Eighties, the tales of Noggin the Nog (a cartoon character that had a flying long boat and a friendly dragon), and subsequently finding sites to recapture the magic of this show, have made it clear to me that these fan sites of old shows are a good thing.
I'm sorry if I came across as rude, it was not intended.
Comp Sci is an odd animal. The teaching of algorithms and numerical analysis were small parts of my degree. As you say, different schools teach it in different ways, and given my options again I would have gone to a different university to do my courses.
That said, in all the years I have been generating solutions for people I have only ever used one mathamatical formula (Boyles Law). I have used may sort algorithms, but after the 4th or 5th decided that the language libraries generally had good enough algorithms. I found that the criteria for getting things to work changed. Where in school it had been "the best wins," in the office it was "what fits in 128 bytes?" or "can it be done in 2 days?" These were facets of computing that cut across the lines, be they Computer Science, Software Engineering, or general purpose hacking.
I find it funny that you think of me as a business person. I suppose to an extent I am, but only so far as I need to find work so that I can eat and buy toys. I love computing. Always have. (Saw my first computer in 1965, took up a good bit of a room and it had less power that a wristwatch does today)
I program for a living. That means understanding not just how to type code in, but how to structure applications to best handle a users needs, to fit within artificial limits (like cost, machine capabilities, user experience), and to make the systems seem simple enough to be able to have people run them.
Computer Science has its place. Understanding that one algorithm is more suitable than another is definitely a desireable skill. Understanding how to build applications when the technology has moved along through many generations, and what you thought once was tough is now just one small part of a hardware call, is also a skill that many never quite catch on to. In 6 or 7 years time the technology you and I are using today will be old and stale. The new stuff will be fast and flashy. It will have every kind of sort you ever heard of and a few more besides as part of some distributed ether library, run on a cluster of machines in Russia, China and India. It will have 3D graphic features to generate component manufacturing instructions for instant prototype component creation on a machine in Dusseldorf, then turn the same thing into metal or ceramics in a machine tool in Indianapolis, or Prague, or Jakarta.
Someone will write that code. It won't be me. If you're lucky it could be you. It will be hard, it will take a long time to come together, it will take changes in technology and thinking. And the people that do it will be masters. But at the end of the day will they be Computer Scientists or just ordinary folks that worked long hours?
Your University courses will give you the ground work to start the process of teaching yourself what you will need in the future. If all you take from your university is the thirst for knowledge then they will have given you all you need. But if they give you algorithms, and concrete views of what should and shouldn't be, then you have lost.
And in closing, remember the one truth: Computers are for people, are used by people, and built by people. People count more than computers.
In my current role I am a CTO for a small startup. I do everything from architecture, coding, pulling cable, making sure the backups are done, creating data table space on the servers, putting fax paper in the fax machine, making sure the refrigerator is full of Coke cans, through buying the systems everyone works on. All has to be within budget and on time. And the investors and clients have to see that.
Over the years I have build ASIC chips, coded parallel processing systems, real-time control systems, database systems, comms systems, yack, yack, yack.
I was invited to go to Rutgers University, here in NJ, a few years ago to discuss Industry/University partnerships. They were very gracious and had several senior faculty members give presentations about their work. They were doing interesting work. One was doing distributed parallel processing, one doing some work on 'data fountains' (what we would call wireless access points.) These were solid projects but each lacked a great deal of what was needed to make them viable solutions in the real world.
I don't doubt you are having a cool time at University, but after 20 years of lay-offs, down sizings, arrogant management, and drop-dead dumb co-workers, I have to say that from my experience you have a few lessons to learn that no University can ever teach you.
I object on every point. Universities are there to make sure they can say they handed out so many BSc's, so many BA's, so many Masters, and so many PhD's. They are not there to teach you how to become good at your chosen profession.
I did a Comp Sci degree. From day one I knew more about computers, programming, getting systems to work, and generally getting around the business that most of the lecturers. And I was not alone.
I found University an appaling waste of my time. I learnt more in the 3 months after I finished that the 3 years there.
Many people left the degree with good passes and still could not tell you how to create a system or determine which way up a keyboard was meant to be placed.
Inspired people should probably do degrees in something else, anything else other than CS. CS does not teach you what you need to know about what you will need in business.
Part of this is inherent. Computer work in the real world is a group affair. Computing at Uni has to be a single persons work. The two are at odds with each other. Further, most universities while they have some top-notch systems, rarely do enough intersting stuff with them for the CS'ers, such that once out of University most people if they go into industry, will end up quickly doing more interesting computing on the PC on their desk.
Computing is not for everyone. I taught computing to business people, years ago, and was surprised that few of them could get their heads around Arrays. Simple stuff to me - essence of life - but to them it was so abstract it didn't mean anything to them.
I did my degree 20 years ago. Today it is just a bad memory. The skills I have today are based on my aptitude for the business and my enthusiasm to create new things. If I had not these qualities when I went to University, there is no way they could have taught them to me.
And in this day and age... Do Art History, it would be much more useful 1 year out than your CS.
Slashdot Mirror
I was quite surprise in doing some OCCAM programming on Transputers, by the level to which you could turn processes into something that ran well concurrently.
Part of the problem we have at the moment is that setting up threads, or forking, is a tedious business and prone to debugging hell. In OCCAM you could write your procedures and where and when you wanted them to run serially or in parallel, put a case statement like structure that was headed with SER or PAR.
Seeing this, it was easy to think of new ways in which you could decompose existing problems and then parallel-ize them to make the most of the hardware.
C++ has been a pile of steaming crap for better than 10 years. Review the success of Java in the period. Cross refernece with the continued us of C to do most (non-windows) systems programming.
As for the Author being the Chair of the standards committee, this just goes to prove we all make mistakes once in a while.
- The IQ difference between a worm and a human is such that it allows the worm to burrow deeply and get out of trouble.
I dunno! I think your Pentium 133 is running slower than most. If todays 'average' PC is a 3GHz screamer that would make your new machine 22 times faster than the old Pentium 133 paper weight.
There has been some improvement in how fast a single instruction operates on the latest Intel chip generations but nothing to get you being 'a few hundred' times quicker.
The Athlons are a different game, AMD did tweak the instruction set to run faster, hence their nominal speed ratings are very different from their actual speed ratings. The nominal rating is what an Intel chip would (surely) have to be clocked at to achieve the throughput. But still not to achieves hundreds of times the performance.
Personally, I need all the speed I can get. I need to get into the top 2 percentile of all Seti@home users... NOW!!! And that means Powwa!! Quad-10GHz processors... mummm!
Possibly.
Could it be that the interceptor developed Alzheimer's?
Blue touch paper lit... Now do what?
I'm not sure our ICBM's are accurate enough for what you want. A Minuteman III had a CEP of about 300 ft. That mattered little if the 500Kt warhead went off. But landing 500lb's of dummy warhead concrete 300ft from a house is not going to do anything but make a loud noise and make a small hole.
But that makes a great project for these guys to get on - improving our ICBM accuracy so we can put concrete blocks in the nose instead of thermonuclear devices. Yes, another way to spend billions!
This is just contorted mumbo-jumbo!
It is either possible to hit something or it is not. Throwing chance into the equation is fence sitting.While the tests have proven that the task of vaporizing an incoming warhead is hard, they have also proven it to be possible
At a tactical level, making sure that the right bits are hit, and not decoys, becomes something that will eventually fall into place. However, that is possible - we do it in other areas of missile interceptions.
What I object to in all of this, is that we will not be attacked by a missile force, therefore this method of defence is useless, and an extreme waste of money. If China or Russia wanted to nuke us with missiles, this system would not shoot down enough. And if North Korea wanted to shoot at us, the return address is easily read. If a terrorist launched a scud missile from a ship in the Carribean, it couldn't be detected or intercepted anyway - it would be coming from the wrong direction and be too low
We had to pretend we were carrying walkie-talkies...
All people are different. Some like symbols, others like patterns, others smells. To say that people who can't remember passwords are stupid, is a sign of limited thinking on your part.
I barely remember peoples names, yet I have met people that have remembered the name of everyone they have ever met.
I once memorized the hex op codes for the 8086 instruction set and could look at a memory dump and generally work out what it did without reference to a manual, or the disassembler. But I can't save multiple passwords in my head for a nanosecond.
There is no one-size-fits-all for human minds.
"Lame-O!"
Error. Password Invalid.
Only seven charaters. Must be 8.
Worked at UPS in New Jersey back in 1993. One particularly ugly morning - there were many ugly days aqt UPS - the automatic password expiration process kicked in. It threw me off my session and demanded that I change my password. I myself was in an ugly mood, so selected as my password "fu*kyou". This seemed appropriate for a place like UPS, at that time. I logged back on and continued to do my work.
Within an hour I was kicked off again, and couldn't log back on. I called systems and they told me the account was disabled and I should talk to my manager. My Manager? I saw the manager and he called, they said they needed his authorization to re-enable my account...
I called systems again. This time they were a bit more willing to talk. It turns out the passwords were held by the UPS mainframe in clear text... and monitored. The sys admin girl seemed a little shocked on accessing my account. I asked, "Was it the password?" "Oh! Yes!" was the reply.
So, security be damned.
I have to say, I have never again used that password!
That's why there's such a wretched stink around here...
I concur. I tried both Solaris 7 and 8 as x86 and they were worse that anything I have ever tried - including OS/2 1.2. There is NO driver support for Solaris. It may be great as a server system with a few proprietary hard drive controllers, but forget doing any desktop stuff with it.
Sun's concept of desktop is also a little warped. I don't think they have ever got their heads around the fact that people unplug their PC's from the mains and wander around with them - still on.
I'm sorry but I think you are being the Nimrod here. With the downturn having taken up the better part of 4 years and .NET having only been a viable option for most people over the last 2 years, there is no way that there are that many people with both legacy language skills (C/C++) and new techniques (C#, VB.NET) in a production environment.
.NET product roll out experience.
.NET if necessary.
This is compounded by your request for product roll out. Few people anywhere have
You are trying to do more with less people, but the combination of skills that you are requesting to make up this deficiency just may be counter intuitive and perhaps even mutually exclusive.
Look at people with Java or C++ skills and say you will train in
Banging a stake in the ground and telling the candidates they must be HERE! Today! is a little presumptuous.
Maybe we could call this.... an appruntecip... no no ... how about an... an "apprenticeship!" Yes! that sounds like a good word.
But they will need 3 years experience before they can start...
Yup, you have to completely uninstall the prior version. The uninstall WILL NOT remove your prior settings, they are retained. It's an act of faith, just Believe! Brothers and Sisters!
Ahem... Sorry for that outburst, I am a survivor of the vi / emacs religious wars and sometimes I get flash backs.....
It could be that you need some serious analysis time in the future if the voices sound like "write or talk"!
Being older, my voices come as if from a telex, as a sort of clacking sound. Or is that my bones?
Ahh, voices in your head...
Does the voice sound like your mother or your father?
Yup! This is probably the way to go. Many other solutions mentioned here will either not work or will dissolve the PCB or its lacquer coating, rendering the electronics useless.
Of course TCTFE is a carcinogen, is mostly banned in all civilized parts of the world and does a number on your brain if you sniff enough of it, but hey, it's a great cleaner.
ICL, that fine (Anglo-)Japanese computer manufacturer - and my first employer - had a whole mainframe they salvaged from a fire. It was taken apart completely, the whole thing then dumped in this cleaner, dried, and re-assembled. The machine ran better and more reliably that any ICL mainframe of the day, though that is not necessarily saying much.
Reggie, now over at MySQL, for the development of the ADO.NET connector. Fine piece of work.
.NET environment and he had the solution I liked best.
I needed to access MySQL in a C# and
Well done dude.
I read the conclusion of this story and have a comment: These were not hobbist computers and parts. Instead they were really what at the time were called Business Systems. Byte in those days was subtitled something like "For Small and Business Systems" making a case for the classification of these computers and their use. I worked in 1981 for a small company selling "Business Systems" (ZR Computers). The $6000AU disk drive was for businesses, not the home computer owner. NASCOM-1's were for Home buyers, as were many smaller Z80 / CP/M systems (Trash 80 and Sharp MZ80's come to mind).
My first job for ZR was to work with a 'network' of 3 Commodore 8096 PET's that used a single Commodore 10MB disk unit. 2 computers used the the IEEE-488 bus to access the Hard drive, the other used the network to communicate with one of the drive connected PET's to get it's data. I called it Master/Slave. But what did I know about political correctness. This drive was a 5000 Pound UK drive unit, at a time when the Pound was worth something (!)
The software I wrote, a machine tool parts catalogue, was billed out at 400 pounds UK (say $800) and took me about 6 months to write. I was paid 3000 pounds / annum, therefore the software cost ZR, even at my stunningly low wage rate, 1100 pounds. Business idiocy is not something new!!
RPN... Somewhere around here is my old Sinclair Scientific (circa 1977). If only the on/off switch still worked. Clive Sinclairs early UK RPN calc. And it was a pretty thing, too, all in white.
But it gets me thinking. I have a roman numeral calculator on the website, could I do an RPN (Reverse Ptolemaic Notation... Poland didn't exist in 200BC) Roman Numeral Calculator? Hmmm! Now there's a project!
II II +
IV
I'm all excited!!!
I don't doubt that there are better programs now, and that some of the shows we watched in years of old were junk, but every so often it is just great to come across something that you loved and makes you feel warm and fuzzy once again. And while not of the Eighties, the tales of Noggin the Nog (a cartoon character that had a flying long boat and a friendly dragon), and subsequently finding sites to recapture the magic of this show, have made it clear to me that these fan sites of old shows are a good thing.
I'm sorry if I came across as rude, it was not intended.
Comp Sci is an odd animal. The teaching of algorithms and numerical analysis were small parts of my degree. As you say, different schools teach it in different ways, and given my options again I would have gone to a different university to do my courses.
That said, in all the years I have been generating solutions for people I have only ever used one mathamatical formula (Boyles Law). I have used may sort algorithms, but after the 4th or 5th decided that the language libraries generally had good enough algorithms. I found that the criteria for getting things to work changed. Where in school it had been "the best wins," in the office it was "what fits in 128 bytes?" or "can it be done in 2 days?" These were facets of computing that cut across the lines, be they Computer Science, Software Engineering, or general purpose hacking.
I find it funny that you think of me as a business person. I suppose to an extent I am, but only so far as I need to find work so that I can eat and buy toys. I love computing. Always have. (Saw my first computer in 1965, took up a good bit of a room and it had less power that a wristwatch does today)
I program for a living. That means understanding not just how to type code in, but how to structure applications to best handle a users needs, to fit within artificial limits (like cost, machine capabilities, user experience), and to make the systems seem simple enough to be able to have people run them.
Computer Science has its place. Understanding that one algorithm is more suitable than another is definitely a desireable skill. Understanding how to build applications when the technology has moved along through many generations, and what you thought once was tough is now just one small part of a hardware call, is also a skill that many never quite catch on to. In 6 or 7 years time the technology you and I are using today will be old and stale. The new stuff will be fast and flashy. It will have every kind of sort you ever heard of and a few more besides as part of some distributed ether library, run on a cluster of machines in Russia, China and India. It will have 3D graphic features to generate component manufacturing instructions for instant prototype component creation on a machine in Dusseldorf, then turn the same thing into metal or ceramics in a machine tool in Indianapolis, or Prague, or Jakarta.
Someone will write that code. It won't be me. If you're lucky it could be you. It will be hard, it will take a long time to come together, it will take changes in technology and thinking. And the people that do it will be masters. But at the end of the day will they be Computer Scientists or just ordinary folks that worked long hours?
Your University courses will give you the ground work to start the process of teaching yourself what you will need in the future. If all you take from your university is the thirst for knowledge then they will have given you all you need. But if they give you algorithms, and concrete views of what should and shouldn't be, then you have lost.
And in closing, remember the one truth:
Computers are for people, are used by people, and built by people. People count more than computers.
In my current role I am a CTO for a small startup. I do everything from architecture, coding, pulling cable, making sure the backups are done, creating data table space on the servers, putting fax paper in the fax machine, making sure the refrigerator is full of Coke cans, through buying the systems everyone works on. All has to be within budget and on time. And the investors and clients have to see that.
Over the years I have build ASIC chips, coded parallel processing systems, real-time control systems, database systems, comms systems, yack, yack, yack.
I was invited to go to Rutgers University, here in NJ, a few years ago to discuss Industry/University partnerships. They were very gracious and had several senior faculty members give presentations about their work. They were doing interesting work. One was doing distributed parallel processing, one doing some work on 'data fountains' (what we would call wireless access points.) These were solid projects but each lacked a great deal of what was needed to make them viable solutions in the real world.
I don't doubt you are having a cool time at University, but after 20 years of lay-offs, down sizings, arrogant management, and drop-dead dumb co-workers, I have to say that from my experience you have a few lessons to learn that no University can ever teach you.
I object on every point. Universities are there to make sure they can say they handed out so many BSc's, so many BA's, so many Masters, and so many PhD's. They are not there to teach you how to become good at your chosen profession.
I did a Comp Sci degree. From day one I knew more about computers, programming, getting systems to work, and generally getting around the business that most of the lecturers. And I was not alone.
I found University an appaling waste of my time. I learnt more in the 3 months after I finished that the 3 years there.
Many people left the degree with good passes and still could not tell you how to create a system or determine which way up a keyboard was meant to be placed.
Inspired people should probably do degrees in something else, anything else other than CS. CS does not teach you what you need to know about what you will need in business.
Part of this is inherent. Computer work in the real world is a group affair. Computing at Uni has to be a single persons work. The two are at odds with each other. Further, most universities while they have some top-notch systems, rarely do enough intersting stuff with them for the CS'ers, such that once out of University most people if they go into industry, will end up quickly doing more interesting computing on the PC on their desk.
Computing is not for everyone. I taught computing to business people, years ago, and was surprised that few of them could get their heads around Arrays. Simple stuff to me - essence of life - but to them it was so abstract it didn't mean anything to them.
I did my degree 20 years ago. Today it is just a bad memory. The skills I have today are based on my aptitude for the business and my enthusiasm to create new things. If I had not these qualities when I went to University, there is no way they could have taught them to me.
And in this day and age... Do Art History, it would be much more useful 1 year out than your CS.