I don't come here for the stories, I come here for the comments. Whatever trolling is going on, there always was somebody with relevant knowledge around.
I have been reading/. since 1998 almost every day (changed userID when I changed e-mail address), loading everything in full at level 0 and summaries at level -1 (not because I love trolls, but because some really interesting AC comments never get modded up.) In the Beta, it just doesn't work like that anymore - that is a big difference with all previous/. restylings.
Filled in the survey, like many others did, it seems we are being ignored. I would rather go back to a monochrome 40 character width screen than to the/. beta.
Too little, too late? You missed the target methinks.
Yeah, that much is obvious. I really expected the PC to tank - after all we already had superior architectures, like MC68000 (obviously) and Z8000 (that would have been flamebait back then) and NS32000 was just around the corner.
It was the first time I witnessed how hype and a brand name could trump quality, especially because IBM had deliberately chosen a crippled processor (8088 was a 8086 with an 8-bit bus) - that was a harsh lesson.
They dictated the architecture which is on desks these days
Indeed, that's what I'm whining about. I guess I'll have to get over it some day.
quickly! must google: job interview site:xkcd.com and post whatever is returned. Seriously what relevance does that have? He doesnt even ask an interview question.
Analogies help to understand something... up to a certain point.
It only illustrates the basic concepts. After that, one has to go beyond the analogy and do the math.
I remember a poster on a door at the math department of my university (parafrazing from memory): "Do not try to visualize a space with more than 3 dimensions. Nobody can do that, trying will just twist your mind. Just use the formulas with the correct number of variables and leave it at that."
I'm probably feeding a troll who made it to the front page here... But let's assume this was an honest question.
Less useful but still useful are command shells. These provide file management mostly. I believe some of them may allow for sending and retrieving email messages.
Shell commands do provide file management, and the moment you realize that almost everything is a file you can start to understand the power of shells. Email messages are nothing but files, they are just on another computer and the shell gives you access to them. The same goes for web pages (wget, links/lynx and even telnet to port 80), video, pictures, sound, invoices, games, even your screen/keyboard/printer/network have an abstraction as a file, whatever...
I used to teach a 3 months Linux system administration class. The first two weeks my students had no GUI at all, just a shell prompt. The idea behind it was: whatever problem you have, once you have shell access you can start to solve it.
After using the system for two weeks in shell mode, I let them install whatever GUI they like. About half my class was on gnome, the other half on KDE and a couple were using OpenBox or something else. The point is: it did not matter which GUI they were on, it was just a tool to open a terminal window, everything was done on the shell anyway. In the beginning of the course, everybody had to use the same distro (SuSE), but by the end of the course the students could use any distro or even BSD or HP/UX (we had an HP Itanium system on loan, because I had called it "Itanic" to one of their sales reps and he wanted me to prove me wrong). For most things this again did not matter, and for the things where it did matter (for example firewalls, boot sequences, virtual file systems) it was interesting. For the final exam they had to install an operating system with mail/SaMBa/Apache server using a distro they had never used before, manually editing the configuration files using vi or whatever text editor they liked most. They also received a spreadsheet with usernames and default passwords which they had to convert to a text file and write a script to create the user accounts with correct privileges from that file (they did not get enough time to click on "add user" 200 times). My students went on and became great sysadmins.
Once you really know how to use the shell, most differences between *nix systems become almost transparent and tedious jobs can be automated easily. So, NO, shells are not "less useful" - shells unlock the true power of your computer.
I agree with you that one should only have kids if one really wants to. Trying to force kids on somebody who does not want them is wrong indeed.
My point was that having kids does not make it impossible to pursue things outside of work. I should have replied to the GP, in stead of to you - my apologies for that.
I have two kids, had to raise them on my own over the past 16 years without any help from their mother since they were 3 and 2 years old. The oldest has now left the house, the youngest is still living with me.
My solution was: work for the money (as programmer and/or system administrator) until 5PM, pick up kids from school and take care of them until 10PM. After 10PM, I studied for my masters (philosophy of science & logic) for my personal enjoyment and as a (rather dubious & risky) investment in my next job. It worked out: I'm currently professor at a university college - teaching maths, information systems & philosophy (and enjoying it, I hope to do this until my students tell me I'm too senile for it).
It is very possible, it just takes some organization and the zeal to do it.
Never regretted having kids: if I got sent back in time I would do it all over again. I agree that "legacy" is worthless: it's about what we do with this life.
My first comp was a TI-99/4A. Learned BASIC and assembly language there. Then on to the C64 for more of the same plus LISP. Then on the the Amiga where I learned the holy language: C
Archimedes > Amiga, Atari ST
I did not buy a TI-99 because it couldn't be programmed in assembly when I was shopping (1981). It had TI Basic and TI cartidges, no third party software. It is only now that I learn that Texas Instruments later (1983) softened their stance on this.
Ultimately, it's Microsoft's fault. They invented (and defined) this whole category in the first place, and any imitation of Microsoft Office will end up suffering the same massive feature bloat and quickly become a slug.
You really believe that Microsoft invented "office software" as a bundle of wordprocessor, spreadsheet & database?
MS Office was introduced in 1990. Forefront's (later bought by Ashton-Tate) Framework was in 1984.
Or maybe all CPUs today are very generalized RISCy architectures with everything taken care of in microcode (or maybe nowadays it's nanocode)? That would make it (comparatively) really easy to do, right?
I agree with most of your points. Of course there is more to it. I apologise for not including the programmers, table designers, network folks, etc... among the non-hype, my bad. On the mixed installations I've been working on, the systems to pump data from OLTP to OLAP were for example most intricate.
The cubes etc... were done by middleware or client software, underneath it everything was relational. The tables were remodelled versions of the production systems: big item table, huge item_detail table, a bunch of reference tables with product codes, customer adresses and hour tables. The ways in which the tables were physically implemented, regarding partitions etc..., was different between OLTP & OLAP, but using the same kind of RDBMS (Oracle, Sybase, PostgreSQL, etc...).
The OLAP queries were not written by programmers. They were generated by management and analyst, using drag & drop in the client software. Manual SQL tuning of those generated queries easily results in an increase of execution speed by factor ten, from 50 hours to well below 5 hours.
I was being provocative in my formulation because the OP was in my opinion flowing a bit too much with the hype. The point I tried to make concerning the original "Ask/. question" was: switching between OLTP & OLAP is just learning some new tools and techniques, which is always a good thing - just don't listen too much to marketing when deciding what your carreer will be (unless if you want to have marketing carreer, then by all means go ahead). There is no "carreer change" from OLTP to OLAP, it is just more of the same: data in, data stored, data retreived, data out, data processed, old data purged or perhaps archived, backups taken.
OLTP: Online Transaction Processing: you buy a ticket, you want it immediatly. The seller types it in the computer and prints your ticket, a database checked if there were free seats and immediatly reserves one. "Immediatly" is the important word, the customer is not waiting.
OLAP: Online Analytical Processing: How many seats from the US to EMEA have been sold by that kind of sellers with such produyt code. Managements wants the results by the end of the month, it is OK if the query runs a couple of hours/days.
Many real life systems contain two databases, one tuned for speed (containing only the current tickets) and one for reference later (containing all tickets sold in the past n months/years). The difference between them is database tuning and SQL tuning. All the rest (such as path to architect: yeah, the more different systems you know, the more chance you will have to design new ones) is hype.
Why not a public database of ALL the known celestial objects? It could be like a web service where you send it parameters, such as coordinate range, magnitude, object type, etc, and a CSV or XML list comes back.
I wonder how many library-of-congresses that is?
Something like Simbad?
The interface is tedious (but powerful), the amount of data and maps/photos is enormous.
Slashdot Mirror
I don't come here for the stories, I come here for the comments. Whatever trolling is going on, there always was somebody with relevant knowledge around.
/. since 1998 almost every day (changed userID when I changed e-mail address), loading everything in full at level 0 and summaries at level -1 (not because I love trolls, but because some really interesting AC comments never get modded up.) In the Beta, it just doesn't work like that anymore - that is a big difference with all previous /. restylings.
/. beta.
I have been reading
Filled in the survey, like many others did, it seems we are being ignored. I would rather go back to a monochrome 40 character width screen than to the
Too little, too late? You missed the target methinks.
Yeah, that much is obvious. I really expected the PC to tank - after all we already had superior architectures, like MC68000 (obviously) and Z8000 (that would have been flamebait back then) and NS32000 was just around the corner.
It was the first time I witnessed how hype and a brand name could trump quality, especially because IBM had deliberately chosen a crippled processor (8088 was a 8086 with an 8-bit bus) - that was a harsh lesson.
They dictated the architecture which is on desks these days
Indeed, that's what I'm whining about. I guess I'll have to get over it some day.
I remember thinking "too little, too late" when IBM launched its x86 line (the IBM 5150 PC with 8088 CPU) in 1981.
Damn, over 30 years later and we're still stuck with a variant of that architecture!
Got bogged down by the rules.
Players don't need to care about the rules. They just say what their character is trying to do and roll a dice if needed.
Only the DM has to know the rules, and he or she is not bound by them.
Also: rules are for beginners. An experienced DM does not need them.
quickly! must google: job interview site:xkcd.com and post whatever is returned. Seriously what relevance does that have? He doesnt even ask an interview question.
Hey, I googled one which is relevant: 1088.
Math is just a flawed model, too.
Well, math as such is not a model, it is used to build models.
But you have a point: the model is always an approximation of what is being modeled.
Analogies help to understand something... up to a certain point.
It only illustrates the basic concepts. After that, one has to go beyond the analogy and do the math.
I remember a poster on a door at the math department of my university (parafrazing from memory): "Do not try to visualize a space with more than 3 dimensions. Nobody can do that, trying will just twist your mind. Just use the formulas with the correct number of variables and leave it at that."
Less useful but still useful are command shells. These provide file management mostly. I believe some of them may allow for sending and retrieving email messages.
Shell commands do provide file management, and the moment you realize that almost everything is a file you can start to understand the power of shells. Email messages are nothing but files, they are just on another computer and the shell gives you access to them. The same goes for web pages (wget, links/lynx and even telnet to port 80), video, pictures, sound, invoices, games, even your screen/keyboard/printer/network have an abstraction as a file, whatever...
I used to teach a 3 months Linux system administration class. The first two weeks my students had no GUI at all, just a shell prompt. The idea behind it was: whatever problem you have, once you have shell access you can start to solve it.
After using the system for two weeks in shell mode, I let them install whatever GUI they like. About half my class was on gnome, the other half on KDE and a couple were using OpenBox or something else. The point is: it did not matter which GUI they were on, it was just a tool to open a terminal window, everything was done on the shell anyway. In the beginning of the course, everybody had to use the same distro (SuSE), but by the end of the course the students could use any distro or even BSD or HP/UX (we had an HP Itanium system on loan, because I had called it "Itanic" to one of their sales reps and he wanted me to prove me wrong). For most things this again did not matter, and for the things where it did matter (for example firewalls, boot sequences, virtual file systems) it was interesting. For the final exam they had to install an operating system with mail/SaMBa/Apache server using a distro they had never used before, manually editing the configuration files using vi or whatever text editor they liked most. They also received a spreadsheet with usernames and default passwords which they had to convert to a text file and write a script to create the user accounts with correct privileges from that file (they did not get enough time to click on "add user" 200 times). My students went on and became great sysadmins.
Once you really know how to use the shell, most differences between *nix systems become almost transparent and tedious jobs can be automated easily. So, NO, shells are not "less useful" - shells unlock the true power of your computer.
I agree with you that one should only have kids if one really wants to. Trying to force kids on somebody who does not want them is wrong indeed.
My point was that having kids does not make it impossible to pursue things outside of work. I should have replied to the GP, in stead of to you - my apologies for that.
Rubbish.
I have two kids, had to raise them on my own over the past 16 years without any help from their mother since they were 3 and 2 years old. The oldest has now left the house, the youngest is still living with me.
My solution was: work for the money (as programmer and/or system administrator) until 5PM, pick up kids from school and take care of them until 10PM. After 10PM, I studied for my masters (philosophy of science & logic) for my personal enjoyment and as a (rather dubious & risky) investment in my next job. It worked out: I'm currently professor at a university college - teaching maths, information systems & philosophy (and enjoying it, I hope to do this until my students tell me I'm too senile for it).
It is very possible, it just takes some organization and the zeal to do it.
Never regretted having kids: if I got sent back in time I would do it all over again. I agree that "legacy" is worthless: it's about what we do with this life.
Amiga > Atari ST
My first comp was a TI-99/4A. Learned BASIC and assembly language there. Then on to the C64 for more of the same plus LISP. Then on the the Amiga where I learned the holy language: C
Archimedes > Amiga, Atari ST
I did not buy a TI-99 because it couldn't be programmed in assembly when I was shopping (1981). It had TI Basic and TI cartidges, no third party software. It is only now that I learn that Texas Instruments later (1983) softened their stance on this.
Ultimately, it's Microsoft's fault. They invented (and defined) this whole category in the first place, and any imitation of Microsoft Office will end up suffering the same massive feature bloat and quickly become a slug.
You really believe that Microsoft invented "office software" as a bundle of wordprocessor, spreadsheet & database?
MS Office was introduced in 1990. Forefront's (later bought by Ashton-Tate) Framework was in 1984.
(And to my surprize, it still exists.)
Or maybe all CPUs today are very generalized RISCy architectures with everything taken care of in microcode (or maybe nowadays it's nanocode)? That would make it (comparatively) really easy to do, right?
Sounds like you are reinventing the Crusoe processor.
I think that was that he ment by tuning. Just the small matter of programming you know..
Absolutely.
I agree with most of your points. Of course there is more to it. I apologise for not including the programmers, table designers, network folks, etc... among the non-hype, my bad. On the mixed installations I've been working on, the systems to pump data from OLTP to OLAP were for example most intricate.
/. question" was: switching between OLTP & OLAP is just learning some new tools and techniques, which is always a good thing - just don't listen too much to marketing when deciding what your carreer will be (unless if you want to have marketing carreer, then by all means go ahead). There is no "carreer change" from OLTP to OLAP, it is just more of the same: data in, data stored, data retreived, data out, data processed, old data purged or perhaps archived, backups taken.
The cubes etc... were done by middleware or client software, underneath it everything was relational. The tables were remodelled versions of the production systems: big item table, huge item_detail table, a bunch of reference tables with product codes, customer adresses and hour tables. The ways in which the tables were physically implemented, regarding partitions etc..., was different between OLTP & OLAP, but using the same kind of RDBMS (Oracle, Sybase, PostgreSQL, etc...).
The OLAP queries were not written by programmers. They were generated by management and analyst, using drag & drop in the client software. Manual SQL tuning of those generated queries easily results in an increase of execution speed by factor ten, from 50 hours to well below 5 hours.
I was being provocative in my formulation because the OP was in my opinion flowing a bit too much with the hype. The point I tried to make concerning the original "Ask
OLTP: Online Transaction Processing: you buy a ticket, you want it immediatly. The seller types it in the computer and prints your ticket, a database checked if there were free seats and immediatly reserves one. "Immediatly" is the important word, the customer is not waiting. OLAP: Online Analytical Processing: How many seats from the US to EMEA have been sold by that kind of sellers with such produyt code. Managements wants the results by the end of the month, it is OK if the query runs a couple of hours/days. Many real life systems contain two databases, one tuned for speed (containing only the current tickets) and one for reference later (containing all tickets sold in the past n months/years). The difference between them is database tuning and SQL tuning. All the rest (such as path to architect: yeah, the more different systems you know, the more chance you will have to design new ones) is hype.
Why not a public database of ALL the known celestial objects? It could be like a web service where you send it parameters, such as coordinate range, magnitude, object type, etc, and a CSV or XML list comes back.
I wonder how many library-of-congresses that is?
Something like Simbad? The interface is tedious (but powerful), the amount of data and maps/photos is enormous.