uhm, you CAN bind a mac addr to an ip with dhcp...
dhcp is convenient to put on all boxes, these days (that aren't switches or routers). simply just enter the mac and ip pair into dhcpd.conf and you're done.
it is by the fear of unemployment, alone, I set my car in motion, it is by the java IDE that programs aquire bugs, the mouse exhibits clicking, the clicking responds to warnings. it is by the fear of unemployment, alone, that I set my car in motion.
basically, we decided we should destroy our own freedoms before anyone else can get a chance to.
it just is that simple. and if we keep people in a false sense of fear, they can be controlled and manipulated to do the Big Man's bidding, whatever and whoever that is, this week.
I don't think at all about terrorists. but I do think about the loss of freedom, almost weekly, now. I know who 'broke' things and I won't listen to their lies ever again. if their lips are moving, they are lying; its very easy, now.
we don't need more 'silicon valleys'; we need more enlightened companies to RESUME the progress toward telecommuting.
instead, we have biatches like the yahoo girl who reverse progress and force people who work in software to get in their cars, burn more fuel, wait and waste time in traffic (twice) and get to work more tired than if they just rolled out of bed and logged in from home.
then, you could let people buy houses where its a cultural match with them and where they can afford it. I'm middle aged and cannot afford to buy a livable house in the bay area and that's just a damned crime, given that my parents who made a fraction of what I make could easily afford a house, kids, etc - all on a SINGLE income!
stop the crazy BS about demanding people commute to work. fix that and we will fix a lot of things all at once.
that was such a classic. most of the movie is classic, but that line sequence was amazing.
I think of it often when 'interviewing' comes up. its just the opposite with me: I can *do* the job, but lately, it seems I can rarely *get* the job. (I suppose that's why that line sticks in my mind).
I usually start out with a data struct (its easy and fast and does show that you know what's going on). then I do some general algorithm or flow discussion (not pseudo code yet but general approach of how I'd solve it and what general alg choices I'd use or not use). I then HOPE that they are happy enough with that so we don't have do dive deeper into code. code is a level of detail that just requires more time. lots of details (any off-by-one issues? did he cover all the corner cases? etc) in the coding but I can walk thru the issues, verbally, saving time and not having to go into gory minutae.
the better (sigh, sorry: older) interviewers usually are ok if we leave things at that level, but the kids insist I go down to the very last line of code and that - sorry to say - really shows more badly on the person asking the question. it shows they don't get it and they need a full answer before they are convinced. more experienced folks don't need to see the ending statement to know the candidate knows this road and has traveled it before.
I use the doctor analogy: suppose you are shopping for a new dentist or doctor or even eye doctor or even shrink. are you going to test or quiz them to see if they are qualified? of course not.
plumbers and electricians? no. you don't quiz them or 'test' them. you hire them, you watch the work and you keep them on or you fire them if you don't like the work. you never TEST them.
so, why is it so acceptable to assume the software guy is lying unless proven otherwise?
in addition to that, do you REALLY think that people can exist at a company for more than a month or so if they can't perform or code or do what the job is about? its not hard to verify if the companies he worked for are real and the time was real there. if he stayed there for a year, do you REALLY think he could not code his way out of a paper bag? really?
its a shame that we are not unionized, as an industry. we let employers walk all OVER us. we have no power to push back or change the game at all. and we show no signs of getting our power back, either.
I would never ask anyone to go thru the whole code base or tarfile.
pick out something, just anything that stands out and ask me about it. or ask me to explain the org of the code modules, why I did things the way I did, what issues I had along the way, what I might do better next time. stuff like that.
those would be very telling/indicating questions. much better than asking to write code on a white board in short time intervals.
you can see my commenting, my indenting, my choice of variable and proc names, my looping style, my vertical spacing style. none of which you'd really get from a 15 minute coding slot in an interview.
finally, stop obsessing on candidates. almost everyone is doing a 'try before you buy' (ie, contract to perm) and not direct perm anymore. if the guy sucks, get rid of him quickly. you'll know in short time if he can code and debug or not. its essentially no risk and you are wasting more time obsessing at the interview stage than is really necessary.
and in fact, it works both ways. I could tell by the questions they ask if they are any good or not, and they should be able to tell VERY quickly if I wrote it or not.
in fact, since its embedded, I could bring the device to the interview, give a demo, show a code change, download, debug/run result. but again, even when I offered, it was never accepted.
employers: please consider opening your minds when folks who don't all dance to the same drummer come to interview. in fact, wouldn't it be refreshing to NOT have to go thru the same old grind, over and over again? if someone asks to demo or explain their code, please consider that as a valid alternative to the 'code me a linked list in 15 minutes while I wait' kind of thing. or 'reverse this string in place'. memorization and solving toy problems in short time slots isn't my bag and its also not necessary for most of the software jobs that people are likely to do.
I have open source code (about 13k lines of c/c++ for embedded platform) and I happen to interview 'badly' when I am put on the spot and asked to code up something in 15 minutes while someone watches me. that's not how I work and I fail horribly at THAT style of interview. note, I am fairly good (not a+ but definitely better than average) at coding in the real world - just NOT in synthetic white-board style interviews.
I've repeatedly asked companies to accept my public source code, review it and ask me about it, if they wish. its what they'd get from me if they hired me and I don't have a problem with them judging me based on my submitted public code. but they just don't download it, they dont' ask, they don't care. they ONLY know the stupid 'script' shit of asking to have a guy code while standing on one leg and twirling a hula hoop on the left arm. well, might as well be, for me; I just don't do 'live/standup' coding. just don't do it and I resent being judged on this artificial metric.
I might have been able to do the stand-up thing in my 20's, but I'm in my 50's now and it does not work for me anymore. I wish employers would get beyond this broken method of interview style. or at least admit that not everyone writes code in 15 minutes while being watched and under pressure. in my years of writing code on the job, never once was there this kind of condition while I worked.
here's your solution: buy a US router and a china router. put them in parallel (on their inputs) and on their outputs, use a local AND. only pass packets that are produced by both and reject all differences.
unless you build CHIPS, you can't build a fully trustable computer anymore. maybe using 30 yr old chips, but not any modern chips.
its easy enough to put firmware and microcode in almost any chip.
would you trust a nic chip? it has firmware and its rom is closed source. cpus? they have closed source 'errata' microcode and even what's deep inside an intel chip is not for you or I to see.
pc's bios? yeah, right. like you can trust that.
basically, nothing is trustable anymore. maybe that 30 yr old trs-80 is, or the atari or amiga or PET computer.
wonder if we'll see a rebirth of those in operation. ebay, here we come!
why? do you REALLY think that the world' 'greatest' spy agency would be so sloppy as to have the mail system (any mail system) log 'route-arounds' that look suspicious?
really? REALLY??
anyone that powerful will have built-in ways to suppress any mail log records. in fact, if you ordered from dell, my GUESS is that dell is in bed with the bad guys and any 'special firmware' that might have to be installed for user X will be done BY dell AT dell, never having to give any indication that wrong-doing happened.
(cough) with china, the backdoors are put there FROM the factory. no trip to the chinese version of NSA needed.
if you trust chinese software or embedded hardware, you are stupid and/or ignorant.
(similar if you trust the US stuff, now, too, sorry to say!)
maybe something good will come from this: the world does not trust as easily anymore. in a way, that can be a good thing; its certainly a maturing thing. the world is growing up and not thinking life is a wonderful disney movie anymore. the world is filled with bad guys and those wearing white are often the worst (so to speak).
mail really doesn't need a 're-invention' in UI. it didn't 10 yrs ago, to be honest. we have understood email for a long time, now.
put me down as one of those that want a TRUE separation of form and function. I first learned UI stuff via motif (yeah, yeah..) and its UIL concept was pretty cool. you could, even as a user, define the UI in one language and form and the back-end code was entirely separate. the back-end would be updated by the programmers but the UI would (or could) stay stable if the end user wanted.
why can't we have that idea for web stuff? and even modern apps?
I have stopped doing updates. no more updates on my phone and no more firefox or thunderbird updates. I'll live with 'older versions' just so that the UI stays the same and won't break on me.
I have test gear for my work bench that has not changed in half a century. the concept of DMMs, decade boxes, scopes, power supplies - all have pretty stable UI's, rarely do they have touch screens and even the ones that have graphic displays don't re-layout their displays ever 3-6mos, on whimsy. test gear does not change its UI and we are happy for it.
I'd like to see must-have apps (mail, web) stay stable in their UI and only get security and bugfix updates on one track; and new features/gui on another. then let people choose the stable track or the update track.
but noooooo. we can't have that. makes too much sense.
how does that work, if you don't have a tv with TUNER but you have a computer monitor and you watch content that is not over cable or sat or over the air?
Slashdot Mirror
uhm, you CAN bind a mac addr to an ip with dhcp...
dhcp is convenient to put on all boxes, these days (that aren't switches or routers). simply just enter the mac and ip pair into dhcpd.conf and you're done.
damn! my fridge has stuck pixels...
(maybe the freezer was to blame?)
item not as described. password salt was actually pepper!
it is by the fear of unemployment, alone, I set my car in motion,
it is by the java IDE that programs aquire bugs,
the mouse exhibits clicking,
the clicking responds to warnings.
it is by the fear of unemployment, alone, that I set my car in motion.
each country can either accept it or reject it through a democratic process.
no, sorry. firstly, countries don't accept or reject things, PEOPLE IN POWER in those countries do.
secondly, well, fuck it, we all know that we lose when it comes to private deals, big money and big spying.
we lost. we will always lose. david won't win against goliath; only in story books does that happen.
yes, we're fucked.
% make me
make: *** No rule to make target `me'. Stop.
I'm sorry, dave, but I just can't do that ...
basically, we decided we should destroy our own freedoms before anyone else can get a chance to.
it just is that simple. and if we keep people in a false sense of fear, they can be controlled and manipulated to do the Big Man's bidding, whatever and whoever that is, this week.
I don't think at all about terrorists. but I do think about the loss of freedom, almost weekly, now. I know who 'broke' things and I won't listen to their lies ever again. if their lips are moving, they are lying; its very easy, now.
leading to socialist-style government regulations that are downright hostile to business
I keep hearing shit like this but you know, there is NOTHING in the US anymore that is hostile toward business. that's BS talk.
consumer rights go down each year, employee rights go down, corporations are now 'people' and freedom = speech = buyable.
if anything, I wish there was a true bit of socialism left in the US, but I see none at all, just fear *about* it existing from ultra conservs.
we don't need more 'silicon valleys'; we need more enlightened companies to RESUME the progress toward telecommuting.
instead, we have biatches like the yahoo girl who reverse progress and force people who work in software to get in their cars, burn more fuel, wait and waste time in traffic (twice) and get to work more tired than if they just rolled out of bed and logged in from home.
then, you could let people buy houses where its a cultural match with them and where they can afford it. I'm middle aged and cannot afford to buy a livable house in the bay area and that's just a damned crime, given that my parents who made a fraction of what I make could easily afford a house, kids, etc - all on a SINGLE income!
stop the crazy BS about demanding people commute to work. fix that and we will fix a lot of things all at once.
that was such a classic. most of the movie is classic, but that line sequence was amazing.
I think of it often when 'interviewing' comes up. its just the opposite with me: I can *do* the job, but lately, it seems I can rarely *get* the job. (I suppose that's why that line sticks in my mind).
I usually start out with a data struct (its easy and fast and does show that you know what's going on). then I do some general algorithm or flow discussion (not pseudo code yet but general approach of how I'd solve it and what general alg choices I'd use or not use). I then HOPE that they are happy enough with that so we don't have do dive deeper into code. code is a level of detail that just requires more time. lots of details (any off-by-one issues? did he cover all the corner cases? etc) in the coding but I can walk thru the issues, verbally, saving time and not having to go into gory minutae.
the better (sigh, sorry: older) interviewers usually are ok if we leave things at that level, but the kids insist I go down to the very last line of code and that - sorry to say - really shows more badly on the person asking the question. it shows they don't get it and they need a full answer before they are convinced. more experienced folks don't need to see the ending statement to know the candidate knows this road and has traveled it before.
I use the doctor analogy: suppose you are shopping for a new dentist or doctor or even eye doctor or even shrink. are you going to test or quiz them to see if they are qualified? of course not.
plumbers and electricians? no. you don't quiz them or 'test' them. you hire them, you watch the work and you keep them on or you fire them if you don't like the work. you never TEST them.
so, why is it so acceptable to assume the software guy is lying unless proven otherwise?
in addition to that, do you REALLY think that people can exist at a company for more than a month or so if they can't perform or code or do what the job is about? its not hard to verify if the companies he worked for are real and the time was real there. if he stayed there for a year, do you REALLY think he could not code his way out of a paper bag? really?
its a shame that we are not unionized, as an industry. we let employers walk all OVER us. we have no power to push back or change the game at all. and we show no signs of getting our power back, either.
I would never ask anyone to go thru the whole code base or tarfile.
pick out something, just anything that stands out and ask me about it. or ask me to explain the org of the code modules, why I did things the way I did, what issues I had along the way, what I might do better next time. stuff like that.
those would be very telling/indicating questions. much better than asking to write code on a white board in short time intervals.
you can see my commenting, my indenting, my choice of variable and proc names, my looping style, my vertical spacing style. none of which you'd really get from a 15 minute coding slot in an interview.
finally, stop obsessing on candidates. almost everyone is doing a 'try before you buy' (ie, contract to perm) and not direct perm anymore. if the guy sucks, get rid of him quickly. you'll know in short time if he can code and debug or not. its essentially no risk and you are wasting more time obsessing at the interview stage than is really necessary.
and in fact, it works both ways. I could tell by the questions they ask if they are any good or not, and they should be able to tell VERY quickly if I wrote it or not.
in fact, since its embedded, I could bring the device to the interview, give a demo, show a code change, download, debug/run result. but again, even when I offered, it was never accepted.
employers: please consider opening your minds when folks who don't all dance to the same drummer come to interview. in fact, wouldn't it be refreshing to NOT have to go thru the same old grind, over and over again? if someone asks to demo or explain their code, please consider that as a valid alternative to the 'code me a linked list in 15 minutes while I wait' kind of thing. or 'reverse this string in place'. memorization and solving toy problems in short time slots isn't my bag and its also not necessary for most of the software jobs that people are likely to do.
that does not work. I tried, many MANY times.
I have open source code (about 13k lines of c/c++ for embedded platform) and I happen to interview 'badly' when I am put on the spot and asked to code up something in 15 minutes while someone watches me. that's not how I work and I fail horribly at THAT style of interview. note, I am fairly good (not a+ but definitely better than average) at coding in the real world - just NOT in synthetic white-board style interviews.
I've repeatedly asked companies to accept my public source code, review it and ask me about it, if they wish. its what they'd get from me if they hired me and I don't have a problem with them judging me based on my submitted public code. but they just don't download it, they dont' ask, they don't care. they ONLY know the stupid 'script' shit of asking to have a guy code while standing on one leg and twirling a hula hoop on the left arm. well, might as well be, for me; I just don't do 'live/standup' coding. just don't do it and I resent being judged on this artificial metric.
I might have been able to do the stand-up thing in my 20's, but I'm in my 50's now and it does not work for me anymore. I wish employers would get beyond this broken method of interview style. or at least admit that not everyone writes code in 15 minutes while being watched and under pressure. in my years of writing code on the job, never once was there this kind of condition while I worked.
here's your solution: buy a US router and a china router. put them in parallel (on their inputs) and on their outputs, use a local AND. only pass packets that are produced by both and reject all differences.
(I'm kidding, but maybe only half kidding..)
I prefer tcp-over-waterbuffalo.
its more robust.
the target customer, here, would be the same one that buys a 'smart tv' instead of using a computer and a tv monitor together.
note: there are more dummies in the world than smart guys. we are a tiny tiny minority in the world.
unless you build CHIPS, you can't build a fully trustable computer anymore. maybe using 30 yr old chips, but not any modern chips.
its easy enough to put firmware and microcode in almost any chip.
would you trust a nic chip? it has firmware and its rom is closed source. cpus? they have closed source 'errata' microcode and even what's deep inside an intel chip is not for you or I to see.
pc's bios? yeah, right. like you can trust that.
basically, nothing is trustable anymore. maybe that 30 yr old trs-80 is, or the atari or amiga or PET computer.
wonder if we'll see a rebirth of those in operation. ebay, here we come!
that is almost guaranteed to be bogus.
why? do you REALLY think that the world' 'greatest' spy agency would be so sloppy as to have the mail system (any mail system) log 'route-arounds' that look suspicious?
really? REALLY??
anyone that powerful will have built-in ways to suppress any mail log records. in fact, if you ordered from dell, my GUESS is that dell is in bed with the bad guys and any 'special firmware' that might have to be installed for user X will be done BY dell AT dell, never having to give any indication that wrong-doing happened.
what they have, then, is a 'garage'. right? its two backdoors right next to each other: the chinese one and the nsa one.
where I come from, 2 back doors right next to each other = "a garage"
and so, we have been letting our citizens install routers with built-in garages... garages big enough to, uhm, drive a truck thru.
(cough) with china, the backdoors are put there FROM the factory. no trip to the chinese version of NSA needed.
if you trust chinese software or embedded hardware, you are stupid and/or ignorant.
(similar if you trust the US stuff, now, too, sorry to say!)
maybe something good will come from this: the world does not trust as easily anymore. in a way, that can be a good thing; its certainly a maturing thing. the world is growing up and not thinking life is a wonderful disney movie anymore. the world is filled with bad guys and those wearing white are often the worst (so to speak).
mail really doesn't need a 're-invention' in UI. it didn't 10 yrs ago, to be honest. we have understood email for a long time, now.
put me down as one of those that want a TRUE separation of form and function. I first learned UI stuff via motif (yeah, yeah..) and its UIL concept was pretty cool. you could, even as a user, define the UI in one language and form and the back-end code was entirely separate. the back-end would be updated by the programmers but the UI would (or could) stay stable if the end user wanted.
why can't we have that idea for web stuff? and even modern apps?
I have stopped doing updates. no more updates on my phone and no more firefox or thunderbird updates. I'll live with 'older versions' just so that the UI stays the same and won't break on me.
I have test gear for my work bench that has not changed in half a century. the concept of DMMs, decade boxes, scopes, power supplies - all have pretty stable UI's, rarely do they have touch screens and even the ones that have graphic displays don't re-layout their displays ever 3-6mos, on whimsy. test gear does not change its UI and we are happy for it.
I'd like to see must-have apps (mail, web) stay stable in their UI and only get security and bugfix updates on one track; and new features/gui on another. then let people choose the stable track or the update track.
but noooooo. we can't have that. makes too much sense.
Then again, Jews are famous for being self-loathing
then again, non-jews are famous for not having clue-1 about non-fiction, in-real-life jews. maybe you think all jews are like woody allen??
how does that work, if you don't have a tv with TUNER but you have a computer monitor and you watch content that is not over cable or sat or over the air?