Storing bitmasks in SQL





  • I think that falls under the catagory of 'student' programmer, given it's from a game mod forum and that users avatar...

     Anyway, would the proper way be to have a group table, with a user-group link table, and a priviledge table with a group-priviledge link table?



  •  Garry's Mod already handles groups and user-group links, and for a fixed set of privileges it seems unnecessary to have a privilege table listing them.



  • Do you think using bitmasks is a bad thing?  I think it would be useful in this case.  It depends on the DB.  Some have Bit types that actually make effiecient use of memory.  Others take up a whole byte or larger for just '1' or '0'.  That's where integers and bit masks come in.  SQL supports bitwise AND.

     Obviously, representing the value as ascii coded binary (or Hex) would be fucktarded.



  • @immibis said:

     Garry's Mod already handles groups and user-group links, and for a fixed set of privileges it seems unnecessary to have a privilege table listing them.


    If there are 50 privileges, before long there will be 51.



  • Someone should pls give zzzaaaaacccckkk the codez …



  • 			<div class="content">
    				<div id="post_message_27706627">
    					
    					<blockquote class="postcontent restore ">&gt; 32 guests viewing???<br>
    

    <font color="#444444" size="1">> Edited:</font>

    > 55 fucking guests
     
    hahaha :)
    			</div>
    		</div>


  • @Cad Delworth said:

    Someone should pls give zzzaaaaacccckkk the code
    You go. Bugmenot failed me and I'm not registering at random forum #523145 to post a single comment. Also someone should tell them about SO; that place is just sad.



  •  I love this bit of advice:

    And if you want even more compact storage, you convert your bitflag to
    hex, and convert it back when you want to read/write to it

    Because a bitflag is (by definition) a collection of bits; and storing the data in hex is really just storing character values, which are stored as a collection of bits...

     

    Why stop there?  Convert your bitflag to hex; then compress it; convert that representation to a bitflag; convert that to hex, then compress that ... pretty soon, you'll be down to just one bit  Ultimate compact storage.



  •  Stating the bleeding obvious, which unfortunately nobody in either this forum or the other has pointed out, is to ask why the flying WTF we need SQL during the game?

    Yes, we'll use an SQL DB to save the user's characteristics, and restore when he/she logs on, but  during the game? No. We use an in-memory lookup table.

     1. How many simultaneous users will there will be? (Say) 1024.(for convenient calculation)

    2. For each one, we need (say) 8 capabilities, so one byte of bitflags, plus (say) an int for player_id (unless we can use table position.)

    3. Total space req'd  1024 x 2 bytes= 2k of RAM

    Conclusion: Bill was right, 640K should be enough for anyone (who knows what they are doing, which is why Windows 7 needs more memory for one user than an IBM mainframe needs for 5,000)



  • @GreyWolf said:

    Stating the bleeding obvious, which unfortunately nobody in either this forum or the other has pointed out, is to ask why the flying WTF we need SQL during the game?

    Wow, gramps, what decade are you from? Games aren't Doom anymore, pretty much all of them use a SQL database during pretty much all of gameplay. (And yeah, I imagine some cacheable items are cached, but probably not as many as you think-- and a permissions lookup isn't going to break the bank, they happen so infrequently.)

    @GreyWolf said:

    Conclusion: Bill was right, 640K should be enough for anyone

    At the risk of being a humorless ass, Bill never fucking said that.

    @GreyWolf said:

    (who knows what they are doing, which is why Windows 7 needs more memory for one user than an IBM mainframe needs for 5,000)

    Windows 7 does 5000 times more stuff than an IBM mainframe.



  •  I'm not arguing, just asking...

    1. What's the path length for a SELECT? In, for sake of argument , MySQL. 400,000 instructions approx at a guess? What's the path length for looking up in an in-memory table? 2? 3? Which method has a better chance of maintaining frame rate?

    2. I know Bill didn't say that, and I apologise to him for misquoting. I was trying to show a contrast in memory efficiency of the in-memory table compared to the SQL method for this case only.

    3. Windows 7 does not do 5000 times as much stuff as a mainframe, because so much of it is neither necessary nor useful.

    4. IBM didn't write Lotus Notes. The clue is in the name. If you want to say that IBM should have strangled Lotus Notes within minutes of acquiring it, and buried it in a crossroads at midnight with a stake through its heart, I would agree totally.



  • @GreyWolf said:

     

     I'm not arguing, just asking...

    You are in the wrong forum then

    @GreyWolf said:

    3. Windows 7 does not do 5000 times as much stuff as a mainframe, because so much of it is neither necessary nor useful.

    Could you like be specific?, pretty much anything that win7 does you can tell it not to do and it is necessary and/or useful for the common user, ergo Microsoft #1 customer



  • Writing on a phone, forgive formatting.

    1. A memory lookup is quicker, but a cached SQL query is nearly as quick. Plus, development time is significantly faster. Plus, since the game logic/AI runs on a different thread than the renderer, the impact on fps is nil. You don't optimize until there is a problem, and 99% of games never hit a problem related to SQL.

    3. Then why is it so popular? Look, I can find 5000 things Windows 7 does that IBM mainframes don't in the hardware compatibility list alone! When I worked with an AS/400 system, it couldn't even talk to printers in postscript or pcl!



  • @blakeyrat said:

    Writing on a phone, forgive formatting.

    1. A memory lookup is quicker, but a cached SQL query is nearly as quick. Plus, development time is significantly faster. Plus, since the game logic/AI runs on a different thread than the renderer, the impact on fps is nil. You don't optimize until there is a problem, and 99% of games never hit a problem related to SQL.

    3. Then why is it so popular? Look, I can find 5000 things Windows 7 does that IBM mainframes don't in the hardware compatibility list alone! When I worked with an AS/400 system, it couldn't even talk to printers in postscript or pcl!

    I can forgive the formatting but not the fallacy!  The fact that Windows is popular in no way proves that it is not bloated and inefficient.  IIf C++ had never had a string concatenation operator, we'd only need a tenth the number of cpu cycles we're currently using.



  • For grins:

    if object_id('tempdb..#permissions') is not null
    drop table #permissions

    create table #permissions ( name nvarchar(32), flags bigint )

    declare @none bigint; set @none = 0
    declare @noclip bigint; set @noclip = power(2,0) --1 -- 2^0
    declare @slap bigint; set @slap = power(2,1) -- 2 -- 2^1
    declare @slay bigint; set @slay = power(2,2) -- 4 -- 2^2
    declare @goto bigint; set @goto = power(2,3) -- 8 -- 2^3

    insert #permissions ( name, flags )
    values ( 'me', @noclip + @slap + @slay + @goto ),
    ( 'you', @noclip ),
    ( 'him', @slap ),
    ( 'her', @none ),
    ( 'she', @noclip + @slay )

    select *,
    case when flags | @noclip = flags then 1 else 0 end as NOCLIP,
    case when flags | @slap = flags then 1 else 0 end as SLAP,
    case when flags | @slay = flags then 1 else 0 end as SLAY,
    case when flags | @goto = flags then 1 else 0 end as [GOTO]

    from #permissions



  • @DaveK said:

    I can forgive the formatting but not the fallacy!  The fact that Windows is popular in no way proves that it is not bloated and inefficient.

    Ok; I concede that. But I've also used an IBM mainframe OS, and it sucked shit. It sucked so much shit we had to pay like $250 per printer to "IBM-ize" it (install a IPDS card and memory) so that the printer could talk to the shitty OS. Unless you bought an IBM-branded printer, but those were just shitty Lexmarks with the $250 worth of hardware pre-installed. IBM stuff sucks. All of it. The mainframes, the Lotus Notes, the terminal emulators... sucks shit.

    @DaveK said:

    IIf C++ had never had a string concatenation operator, we'd only need a tenth the number of cpu cycles we're currently using.

    What.


  • Discourse touched me in a no-no place

    @blakeyrat said:

    @DaveK said:
    I can forgive the formatting but not the fallacy!  The fact that Windows is popular in no way proves that it is not bloated and inefficient.

    Ok; I concede that. But I've also used an IBM mainframe OS, and it sucked shit. It sucked so much shit we had to pay like $250 per printer to "IBM-ize" it (install a IPDS card and memory) so that the printer could talk to the shitty OS. Unless you bought an IBM-branded printer, but those were just shitty Lexmarks with the $250 worth of hardware pre-installed. IBM stuff sucks. All of it. The mainframes, the Lotus Notes, the terminal emulators... sucks shit.

    AS/400's are excellent sources of cooling fans (with which to build elaborate ventilation systems to blow farts at coworkers) and scrap magnesium (Indeed, I am aware of no other reliable source of magnesium in computing equipment up until a couple companies started building laptop frames out of it about 10 years ago)



  • Erm … an AS/400 doesn't count as a mainframe, technically: it's just an overpriced Unix box (unless you're running RPG on it, God help you!). Or did you mean a real mainframe, like, say, a 370/158?


  • Discourse touched me in a no-no place

    @Cad Delworth said:

    Or did you mean a real mainframe, like, say, a 370/158?
    Also a handy source of giant-assed air movers. Not so much the magnesium, though.

     

    Of course, these days the descendents of the mighty 370 are ALSO overpriced Unix boxen:

    and the descendents of THOSE are just glorified blade cabinets accepting x86, x64 and POWER modules. How the mighty have fallen.



  • @Cad Delworth said:

    Erm … an AS/400 doesn't count as a mainframe, technically: it's just an overpriced Unix box (unless you're running RPG on it, God help you!). Or did you mean a real mainframe, like, say, a 370/158?

    Here's what I know:

    1) Our IBM RPG programmer called it an AS/400 mainframe. (Also he programmed in RPG.)
    2) It certainly *wasn't* a PC. (PCs can do things like, say, talk to printers. And interoperate with other systems, including mainframes, in more sensible ways than just shitting out HL7 messages.)
    3) It was in a huge black and red box, and extremely expensive.
    4) We used it as little as possible. By the time I left the hospital, the Phramacy system and ER system were finally migrated off that beastmonster.
    5) It failed just as much as a Windows server would have-- more if you consider that we could have bought 5 beefy Windows servers for the same price and have failover. In particular, if you didn't erase a backup tape before putting it in GOD HELP YOU! It would puke all over itself around 1:00 AM and then it's pagertime!
    6) At no point during my (extensive) use of this machine did I stop to say "this was a sensible design decision" or "those IBM engineers really know what they're doing!"

    So I apologize if it turns out an *actual* mainframe from IBM is a golden land of sun and roses which instantly solves all problems ever, and is well-worth the cost. But somehow? I'm doubting that's the case... I've yet to see a single product from IBM, hardware *or* software (except ViaVoice) that wasn't a total ball of shit.



  • Interesting, blakester: maybe yours was a later model of AS/400 than the ones I knew. When I worked on actual mainframes like the 370 series, the ones that were the size of several wardrobes, we referred to the AS/400s (as IBM did) as 'minicomputers' or 'minis' for short—now that's a term you rarely hear nowadays! The AS/400s that I saw were maybe the size of a washing machine.


    Note that I wasn't defending IBM in any way, just taking issue with your description of an AS/400 as a mainframe. :) I'm sure you would be even more unimpressed with actual IBM mainframes of that era, which had about the processing power (or maybe less?) of a modern mobile telephone, and maybe even less disk storage capacity than said telephone. Oh, and the mainframe peripherals were even MORE 'proprietary/locked-in'— and MUCH more expensive than—those for an AS/400.



  • @Cad Delworth said:

    Interesting, blakester: maybe yours was a later model of AS/400 than the ones I knew. When I worked on actual mainframes like the 370 series, the ones that were the size of several wardrobes, we referred to the AS/400s (as IBM did) as 'minicomputers' or 'minis' for short—now that's a term you rarely hear nowadays! The AS/400s that I saw were maybe the size of a washing machine

    The very concept of computers being categorized by physical size is such a wrong wrongosity that it boggles my mind. I don't go back there much, but I hear-tell they bought a second one and it shipped as a 16U (or so) rack server.

    I forgot to mention another detail that blew my mind when I learned it: every night, while preparing for backups, the entire system was down for about 20 minutes. *20 minutes of downtime a DAY* And everybody there in their IBM brain-haze thought that was perfectly acceptable and normal. For an ER room. To have 20 minutes of downtime a DAY. (That's presuming the download tape was blank, of course, if it ran out of space on the tape it could be down until someone manually logged in and fixed it.) This was in 2002.

    The instant, the exact millisecond, that viable PC server OSes existed, all of those machines should have been kicked to the curb. I'd much rather deal with NT4 or OS/2 or even early Linux.



  • sigh I mentioned the physical size of the boxes for IDENTIFICATION PURPOSES ONLY in an attempt to make it easier for you to know which boxes I was referring to. (Blimey! Do you always jump to such incorrect assumptions?) The term 'minicomputer' referred to a class or category of computers, if you like. Nothing to do with their physical size, bub.


    And FWIW, I'm talking about boxes I encountered in the 1980s, not 2002. PCs barely existed in those days (the PC-AT was leading-edge tech. then: ask your granddaddy), and PC servers were still many years away.



  • @Cad Delworth said:

    Blimey! Do you always jump to such incorrect assumptions?

    What are you saying, I'm a kitten huffer?



  • @blakeyrat said:

    I forgot to mention another detail that blew my mind when I learned it: every night, while preparing for backups, the entire system was down for about 20 minutes. 20 minutes of downtime a DAY And everybody there in their IBM brain-haze thought that was perfectly acceptable and normal. For an ER room. To have 20 minutes of downtime a DAY.
     

    Up till early 2007 our entire division was run on an AS/400.  Everything from production to purchasing, lab, shipping, and everything in between.  I  remember the AS/400 as being not quite the width of a refrigerator and about half as tall.  Every night the system went down for 2-3 hours for "backup" usually at the end of the night shift.  This is a manufacturing operation, running 24/7 and production workers had to enter data in order to keep inventories of raw materials and finished goods up to date.   Nowhere near as critical as an ER, but a major pain in the ass for the people who came in every morning and had to put in all the stuff that couldn't be done the night before.  The AS/400 system was accessed using a Windows client that was literaly just a text mode window, complete with green text on a black background.  No mouse control at all.  Everything was done with numbered menus and using the tab key to move around fields.  As retarded as that sounds,  it was simple and straight forward and worked quite well.  Also, once you learned your way around the menus you could "type ahead" by entering the approriate numbers quickly.

    In 2004 the decision was made to switch to JD Edwards (owned by Oracle) software running on PC servers.  After 3 years of working on transitioning to the new system we were bought by a much larger company which uses SAP, which is a gigantic ball of WTF all by itself.

     



  • @blakeyrat said:

    every night, while preparing for backups, the entire system was down for about 20 minutes. 20 minutes of downtime a DAY And everybody there in their IBM brain-haze thought that was perfectly acceptable and normal. This was in 2002.
     

    That was one incompetent shop, blakey. In 1982, when I was teaching best practice in IBM mainframe IT, I was teaching my students that IBM systems then were capable of multi-year non-stop operation with 100% availability (no service interruptions of any kind). IBM's own internal systems were doing it, and we were teaching our customers how to do it. However, as my then boss put it, "you spend a day with some customers and they move a mile, others move an inch". I was tasked with assessing my customers, and spending my time with those who would move farthest. Obviously yours was one of the ones still living in the 70's, thrirty years in the past. I congratulate you on your escape from their dead hands.



  • @GreyWolf said:

    That was one incompetent shop, blakey. In 1982, when I was teaching best practice in IBM mainframe IT, I was teaching my students that IBM systems then were capable of multi-year non-stop operation with 100% availability (no service interruptions of any kind).

    ... Unless the backup tape was full, then it would shit all over itself.

    What about El Heffe's company, that had the exact same problem with the exact same hardware? Isn't that evidence that the problem was at least somewhat widespread? Why didn't IBM fix it at the source?

    @GreyWolf said:

    However, as my then boss put it, "you spend a day with some customers and they move a mile, others move an inch". I was tasked with assessing my customers, and spending my time with those who would move farthest.

    Well, gee whiz Mr. Wizard, do you think it ever occurred to IBM to make their systems easier to use? I'm sorry our small country hospital couldn't afford your $400 an hour consulting services to make your server work the way all servers should fucking work out of the box by fucking default.

    But, I guess IBM's business plan has always been "sell defective hardware/software, then charge through the nose for consultants to 'fix' it." Which is surely a noble, noble company to work for, not at all sleazy as hell! Of course, even the consultant "fixed" product is still worse than the competitor's product out-of-the-box. (Talking about Lotus Notes again) But don't worry about that! IBM'll just wine and dine the C*O and get the contract signed! After all, they'd never be able to sell Lotus Notes in a billion years if they sold it to people who actually had to use email to actually do their fucking job.

    Look, us west coast technology companies might not wear expensive dark suits, but give me Microsoft even at their absolutely worst over IBM at their best... any day of the fucking week.

    @GreyWolf said:

    Obviously yours was one of the ones still living in the 70's, thrirty years in the past.

    I guess when your server vendor is living in the 70s, you really don't have much of a choice but to live in the 70s too, right? The bigger question is, why the fuck are you defending the fact that it's 2010 and IBM is still living in the fucking 70s? Christ, the only actually innovative product they have is a computer that plays fucking Jeopardy (and people thought Microsoft's Songsmith was useless), and they don't fucking sell the damned thing. Everything else is the exact same product they sold in 1978 (well, ok, let's be generous: 1995) with a bad paint job to make it look like it might kind of not be crap.

    I don't understand how anybody can defend the way IBM does business. They just do not care.


  • Discourse touched me in a no-no place

    @blakeyrat said:

    I guess when your server vendor is living in the 70s, you really don't have much of a choice but to live in the 70s too, right? The bigger question is, why the fuck are you defending the fact that it's 2010 and IBM is still living in the fucking 70s? Christ, the only actually innovative product they have is a computer that plays fucking Jeopardy (and people thought Microsoft's Songsmith was useless), and they don't fucking sell the damned thing. Everything else is the exact same product they sold in 1978 (well, ok, let's be generous: 1995) with a bad paint job to make it look like it might kind of not be crap.
    As I kind-of pointed out, IBM doesn't sell actual honest-to-god mainframes anymore. They sell blade systems that can be used to emulate a mainframe, but are more suited to running modern workloads. They're still enormous piles of proprietary horseshit, but at least it's modern new-agey horseshit, not fucking coprolites.

     



  • @Weng said:

    @blakeyrat said:

    I guess when your server vendor is living in the 70s, you really don't have much of a choice but to live in the 70s too, right? The bigger question is, why the fuck are you defending the fact that it's 2010 and IBM is still living in the fucking 70s? Christ, the only actually innovative product they have is a computer that plays fucking Jeopardy (and people thought Microsoft's Songsmith was useless), and they don't fucking sell the damned thing. Everything else is the exact same product they sold in 1978 (well, ok, let's be generous: 1995) with a bad paint job to make it look like it might kind of not be crap.
    As I kind-of pointed out, IBM doesn't sell actual honest-to-god mainframes anymore. They sell blade systems that can be used to emulate a mainframe, but are more suited to running modern workloads. They're still enormous piles of proprietary horseshit, but at least it's modern new-agey horseshit, not fucking coprolites.

     

    I guess that's progress.



  • @blakeyrat said:

    @GreyWolf said:
    That was one incompetent shop, blakey. In 1982, when I was teaching best practice in IBM mainframe IT, I was teaching my students that IBM systems then were capable of multi-year non-stop operation with 100% availability (no service interruptions of any kind).
    ... Unless the backup tape was full, then it would shit all over itself.

    What about El Heffe's company, that had the exact same problem with the exact same hardware? Isn't that evidence that the problem was at least somewhat widespread? Why didn't IBM fix it at the source?

    Thirded here.  Our AS/400 is down for an hour every night for backups.  The guys on the IBM side know that the system can do an on-line backup, but they think it's better to bring it down.  As an added WTF, the backup job is scheduled when the interactive user is logged out of the physical console.  This means that if no-one at the site remembers, backups don't happen.

    BTW, all of our AS/400 code is RPG.  I once asked them if there was any way they could communicate with the outside world other than email or FTP and they said "no".



  • @blakeyrat said:

    3) It was in a huge black and red box, and extremely expensive.
    I initially misread that as 'explosive'. Though given the magnesium I don't expect I'm too far off the mark.



  •  Hey blakey,

     

    This is nothing to do with how IBM does business and everything to do with typical data centre ignorance creating operational WTFs. And I am not trying to defend IBM, just trying to inject some actual facts into the discussion.

    Does it occur to you that backup is a process that needs some planning and forethought? In any OS or HW? Making a backup process that works out-of-the-box is guaranteed to be a bloated, wasteful, slow, wrongly-timed, disruptive process - that quite likely backs up too much and at the wrong time, therefore cannot do a worthwhile restore. The backup process has to be designed by the customer and integrated into their apps, or it's pointless.

    In 1982 IBM was teaching customers Europe-wide to do the actual backup in the app (so it could backup what was actually needed) to disk so it could run unattended, and was not dependent on the tape or tape-drive We would do a file copy to tape at some other time, so as to not slow the app, and to create a backup that was actually restorable. There has been no need to stop the app to do backups since the early 90's. Customer who had a clue never paid IBM to do stuff for them, only the poorly-managed ones did.

     It occurs to me that you in the US had different economic circumstances and wage structure and therefore did not automate until much later than Europe did. In our neck of the woods, it became cheaper to use disk space than mount a tape in 1980. By 1984, one of my customers only manned the data centre during day shift and then only to answer phones.Everything was automated.

    @Jaime said:

    The guys on the IBM side know that the system can do an on-line backup, but they think it's better to bring it down.
    Ignorant oafs. Presumably young squirts recruited since the people who actually knew stuff were made redundant in the early 90's.



  • @GreyWolf said:

     Hey blakey,

    Hey.

    @GreyWolf said:

    This is nothing to do with how IBM does business and everything to do with typical data centre ignorance creating operational WTFs. And I am not trying to defend IBM, just trying to inject some actual facts into the discussion.

    But but... this is DailyWTF!

    @GreyWolf said:

    Does it occur to you that backup is a process that needs some planning and forethought? In any OS or HW?

    Yes. I mean, if nothing else, you need some place to physically put the data. Of course, if you're a home user, you can use a service like Mozy or DropBox with pretty much zero "planning and foresight", but that doesn't help in a hospital with their HIPAA rules.

    @GreyWolf said:

    Making a backup process that works out-of-the-box is guaranteed to be a bloated, wasteful, slow, wrongly-timed, disruptive process - that quite likely backs up too much and at the wrong time, therefore cannot do a worthwhile restore.

    We need a caption like in that South Park Scientology episode: this is what IBM users genuinely believe!

    @GreyWolf said:

    The backup process has to be designed by the customer and integrated into their apps, or it's pointless.

    Yeah, because it wouldn't be IBM if it wasn't over-engineered and over-thought to fuck and back! Obviously you can't install an IBM server without paying through the ass for 500 hours of $300-and-hour consultants to provide you with 800-page documents in impenetrable gobbledygook!

    Hey, let me hit you with this little logic-bomb: how come our Windows servers (with the same administrator) didn't have the same problem? Do you think our system administrator had some kind of Jekyll and Hyde formula, and whenever he dealt with the IBM hardware he became a bone-throwing caveman? Because, according to you, the only reason our backup required downtime is because he wasn't using enough "planning and foresight" right? But somehow he had enough "planning and foresight" to get the backups on the Windows servers working like a champ?

    So what's the difference? Could it be that *gasp* Windows is better-designed than the AS/400? Designed so well that by default it can complete backups without downtime? Even without the same level of "planning and foresight"?! WOW it's almost as if Windows server is a... gasp... SUPERIOR PRODUCT!

    @GreyWolf said:

    It occurs to me that you in the US had different economic circumstances and wage structure and therefore did not automate until much later than Europe did.

    Oh yeah, obviously the US is waaay behind in IT. That explains why the Internet was created in Japan, and all the major computer OSes right now are made in Russia, and how IBM, Microsoft, Apple, Amazon, Ebay, Google, etc etc are all headquartered in Brazil. I'm glad that I'm talking with someone not-at-all delusional.

    @GreyWolf said:

    @Jaime said:
    The guys on the IBM side know that the system can do an on-line backup, but they think it's better to bring it down.
    Ignorant oafs. Presumably young squirts recruited since the people who actually knew stuff were made redundant in the early 90's.

    Kids these days! GET OFF MY LAWN!

    Look, just admit IBM sells expensive, broken, crap. As scheme to drive sales of their higher-margin consulting services. This is what IBM actually does to make money! It's sleazy, and nobody should stand for it, but since all these broken crap IBM systems got grandfathered-in, and companies are too scared shitless to switch, they somehow make a go at it.



  • @blakeyrat said:

    I've yet to see a single product from IBM, hardware or software (except ViaVoice) that wasn't a total ball of shit.

    Model M keyboard.

    (And the Model F and beam spring terminal boards, but I can only speak from personal experience of using a Unicomp Spacesaver, the current generation of the venerable Model M with all 104/105 keys and a slimmer case. However, they do put gold glitter in the keycap plastic mix and call it "metallic", which is nuts. The jet black Model M13 is luscious, but I want my full 105 keys thankyouverymuch.

    (I must have used actual IBM/Lexmark Model Ms in the past when I worked on PS/2 towers, but I forget now, and mechanical keyboards were still part of my life then, where I thought nothing of them, before I drifted into the mushy world of the rubberdome.)

    What strikes me about IBM, is that I don't really know what they do any more. They're famous for making the best keyboards ever made (unless you feel that accolade belongs to Topre), except they spun that off into Lexmark. They've divested their hard drive range, and they've thrown away their renowned laptop range. It's sad that a company with so many famous products has discarded them all and just clung to Lotus Notes. I also read that they made an absurdly high resolution TFT (something like 3200x2400 in 20" but I really forget now and a chat log search is finding nothing) -- doubt that's still available either.

    Obviously IBM are still doing something, but it just feels like they're fading to nothing.



  • @Daniel Beardsmore said:

    @blakeyrat said:
    I've yet to see a single product from IBM, hardware or software (except ViaVoice) that wasn't a total ball of shit.

    Model M keyboard.

    A) IBM doesn't make the Model M any more, and hasn't for a generation.

    B) I've tried Model Ms and I don't get why geeks jizz all over it. The thing was louder than a jet engine at take-off. If I didn't break my own fingers trying to depress the nasty key action, my co-workers would have broken them for me after hearing the fucking thing. Pretty much the only thing it has going for it is that it lasts a long time-- well, whoop-de-shit.



  • @blakeyrat said:

    A) IBM doesn't make the Model M any more, and hasn't for a generation.

    You said "wasn't a total ball of shit", past tense, indicating that you were including anything IBM ever made, not just current generation products. Therefore, the Model F and M keyboard families count. And although IBM don't sell them now, the production plant is still in operation selling brand new Model Ms, just branded Unicomp Customizer 101/102.

    @blakeyrat said:

    B) I've tried Model Ms and I don't get why geeks jizz all over it. The thing was louder than a jet engine at take-off. If I didn't break my own fingers trying to depress the nasty key action, my co-workers would have broken them for me after hearing the fucking thing. Pretty much the only thing it has going for it is that it lasts a long time-- well, whoop-de-shit.

    I grant you, keyboards are a personal preference, and the Model M is one of the stiffest made. However, you're going to struggle to find such a perfect "well-oiled" tactile response out of any other switch technology. The Topre capacitive switch is the only other contender, but it's a very light and very quiet (and very expensive) switch, and many people actually want audible feedback and a decent amount of resistance.

    Cherry's quietest full travel mechanical is the MX brown, tactile but non-clicky, and the tactile cam gives both the feel and sound a scraping quality. Anything else with a quality feel is going to be noisy (especially current generation classic ALPS (Type I Simplified, AKA Fukkas), oh boy could they wake the dead) unless you want linear (Cherry MX blacks and reds) or want to pay the premium atop a premium for Topres.

    When I first got my Unicomp, I was saddened to discover that the stiffness was just as excessive as a Model M-owning friend said they were (she has them boxed away and can't use them), but after a little bit of acclimatisation I can't deny the superiority of the switch technology. And yes, they're noisy, but that's part of the fun! We want keyboards to make a lot of noise! The sound from a Model M is quite intentional as is the sound from all clicky switches. That said, at work I've restricted myself to a Cherry MX brown as it's the quietest mechanical around, and it still has a loud metallic ring to it at that.

    Next up, I'm getting a clicky (white) Fukka for comparison, as my only other "proper" ALPS is a complicated blue ALPS 286 keyboard, a completely obsolete switch type.



  • @Daniel Beardsmore said:

    the Model M is one of the stiffest made. However, you're going to struggle to find such a perfect "well-oiled" tactile response out of any other switch technology.
     

    I don't get it. You mean to say all the other keyboard keys are not ""well-oiled"" and poor tactile response? I don't even know what "well-oiled" is supposed to mean. Especially since you use it in conjunction with noise and stiffness, which contradict eachother quite directly.

    Because if that's what you're saying then I don't get it. I'm currently typing on a Logitech Internet 350, and, just like my home Logitech, it's soft yet firm, thus perfect.

    @Daniel Beardsmore said:

    after a little bit of acclimatisation I can't deny the superiority of the switch technology

    Are you... a keyboardiophile? You make it sound lke keyboards are advanced tech.

    @Daniel Beardsmore said:

    And yes, they're noisy, but that's part of the fun! We want keyboards to make a lot of noise

    What?! WE DO NOT. Like women, keyboards should be seen and not heard.



  • @dhromed said:

    @Daniel Beardsmore said:
    And yes, they're noisy, but that's part of the fun! We want keyboards to make a lot of noise

    What?! WE DO NOT. Like women, keyboards should be seen and not heard.



    When your PHB thinks you do nothing all day because you work quietly, a loud keyboard is a definite asset, saves you from a lot of nagging/interruptions/other grief.



  • Long ramble, but you did ask me, so …

    @dhromed said:

    I don't get it. You mean to say all the other keyboard keys are not ""well-oiled"" and poor tactile response? I don't even know what "well-oiled" is supposed to mean. Especially since you use it in conjunction with noise and stiffness, which contradict eachother quite directly.

    It's hard to explain succinctly. The noise is intentional -- IBM, Acer, ALPS and Cherry have all developed keyswitches that intentionally click when pressed.

    You want some stiffness in keys so that you don't type wrong keys if your fingers stray, and to give your fingers strong feedback.

    Tactile feedback is both pleasant and functional; if you think back to terminals and 80s home micros with linear mechanical switches you'll understand. Rubberdomes including scissors give you a tactile feel that means nothing as it doesn't correspond to the point where the key actuates (around half travel with a Cherry).

    @dhromed said:

    Because if that's what you're saying then I don't get it. I'm currently typing on a Logitech Internet 350, and, just like my home Logitech, it's soft yet firm, thus perfect.

    Rubberdomes vary a lot in quality and feel, but they're not timeless -- if you get a good one and it dies, good luck replacing it with anything that feels the same. Even Cherry MX switches now are not as good as they used to be, it's claimed, but if FILCO keyboards disappeared, and my keyboard broke, I could switch to whatever company of the day was selling plate-mounted Cherry MX brown and blue keyboards and get something very similar.

    @dhromed said:

    You make it sound lke keyboards are advanced tech.

    Old ALPS switches contained 13 parts -- 13 carefully assembled parts inside every key! They were simplified to 7, and some people still prefer the complicated switches. I'll find out sooner or later, as I'll be comparing a brand new Fukka (Type I Simplified) to an old complicated switch board.

    Rubberdomes provide the tactile response with a dome of rubber, but getting a smooth, precise, clean tactile response out of a mechanical assembly is more difficult and the buckling spring works the best. The tactile point is very smooth, the onset force low (ALPS keys need a hefty push to get them moving) and the overall force distributed well over the whole downstroke.

    Yes most people will consider me mad for liking keyboards, but they all knew I was mad anyway.

    @dhromed said:

    What?! WE DO NOT. Like women, keyboards should be seen and not heard.

    Enough people do that it's still a viable business selling keyboards that are intentionally noisy, although Cherry MX switches are leading by far over the others, perhaps in part because anyone wanting buckling springs just picks up a genuine Model M …



  • @Daniel Beardsmore said:

    Obviously IBM are still doing something, but it just feels like they're fading to nothing.
     

    The "something" is multi-million dollar mainframe computers.



  • @Daniel Beardsmore said:

    It's hard to explain succinctly. The noise is intentional -- IBM, Acer, ALPS and Cherry have all developed keyswitches that intentionally click when pressed.

    You want some stiffness in keys so that you don't type wrong keys if your fingers stray, and to give your fingers strong feedback.

    Yeah, because it's not a real man's keyboard unless your fingers THROB at the end of the day. And make sure you have your Chiropractor on speed-dial because you won't be able to do the full 10 digits with your fingers twisted and broken. You should also hang a pair of these from the front.

    Edit: Wait, WTF. How does "hard to press" == "fingers don't stray"? Does that make any sense?

    @Daniel Beardsmore said:

    Rubberdomes vary a lot in quality and feel, but they're not timeless -- if you get a good one and it dies, good luck replacing it with anything that feels the same.

    If you find a good one, they're like $20. Just buy fucking three of them. STILL cheaper than a Model-M. (And, BTW, that's exactly what I did with my beloved Microsoft Comfort Curve 2000. Seriously, it's cheap, liquid-invulnerable, and works great.)

    In short, a keyboard breaking is not that big of a fucking deal.

    @Daniel Beardsmore said:

    Yes most people will consider me mad for liking keyboards, but they all knew I was mad anyway.

    I guess everybody needs a hobby... and yours is so pathetic it makes my "getting level 50 in all four classes on Global Agenda" look productive as hell.



  • Wow. It's like I've sliced into your body with a sabre and all the gallons of hate have come gushing out like floodwater.

    If something as harmless as a computer keyboard riles you to that degree, I would kindly suggest that it's not me who has the problem.



  •  I love keyboards with a good click-clack sound to them, unless when you are in a office with other people who also have such keyboards



  • @blakeyrat said:

    If you find a good one, they're like $20. Just buy fucking three of them. STILL cheaper than a Model-M.
    Yes, but can you bludgeon someone to death with your flimsy plastic keyboard? I think not! Properly tossed a Model M can kill a man at 10 paces. When the zombie apocalypse comes we'll see who's better off.



  • @DOA said:

    Properly tossed a Model M can kill a man at 10 paces. When the zombie apocalypse comes we'll see who's better off.

    Hmm this raises two questions.

    1) If a Model M can kill a man, what good will it be against the undead?

    2) Once you throw your Model M into the crowd of zombies (remember, its zombies - plural) how will you retrieve it to make the next strike? (perhaps having a bandolier full of cheaper keyboards may be the better approach for defending against crowds).

    And changing the subject. My all time favorite zombie movie is Fido



  • @DOA said:

    @blakeyrat said:
    If you find a good one, they're like $20. Just buy fucking three of them. STILL cheaper than a Model-M.
    Yes, but can you bludgeon someone to death with your flimsy plastic keyboard? I think not! Properly tossed a Model M can kill a man at 10 paces. When the zombie apocalypse comes we'll see who's better off.

    But remember, the ComfortCurve is liquid-proof. Once you wipe the blood and gore and bits of brains off, it'll work as well as ever!



  • @OzPeter said:

    2) Once you throw your Model M into the crowd of zombies (remember, its zombies - plural) how will you retrieve it to make the next strike? (perhaps having a bandolier full of cheaper keyboards may be the better approach for defending against crowds).
    What a stupid question. You just pull it back with the cord.



  • @ender said:

    @OzPeter said:
    2) Once you throw your Model M into the crowd of zombies (remember, its zombies - plural) how will you retrieve it to make the next strike? (perhaps having a bandolier full of cheaper keyboards may be the better approach for defending against crowds).
    What a stupid question. You just pull it back with the cord.
    It was not a stupid question. Because if the zombies are close enough that you can hang onto the cord and still hit them, then you are about to be introduced to a new way of living (well technically non-living)



  • @ender said:

    @OzPeter said:
    2) Once you throw your Model M into the crowd of zombies (remember, its zombies - plural) how will you retrieve it to make the next strike? (perhaps having a bandolier full of cheaper keyboards may be the better approach for defending against crowds).
    What a stupid question. You just pull it back with the cord.

    Watch out for the detachable cord models. Aim for one with drainage holes and a fixed cord.


Log in to reply
 

Looks like your connection to What the Daily WTF? was lost, please wait while we try to reconnect.