00:00:00 --- log: started forth/06.08.21 00:15:33 morning 00:43:08 --- quit: uiuiuiu (sterling.freenode.net irc.freenode.net) 00:43:09 --- quit: Quartus (sterling.freenode.net irc.freenode.net) 00:43:09 --- quit: michaelw (sterling.freenode.net irc.freenode.net) 00:43:09 --- quit: cods (sterling.freenode.net irc.freenode.net) 00:43:09 --- quit: madwork (sterling.freenode.net irc.freenode.net) 00:43:09 --- quit: ccfg (sterling.freenode.net irc.freenode.net) 00:43:11 --- quit: madgarden (sterling.freenode.net irc.freenode.net) 00:43:11 --- quit: Quiznos (sterling.freenode.net irc.freenode.net) 00:43:11 --- quit: fission (sterling.freenode.net irc.freenode.net) 00:43:12 --- quit: astecp (sterling.freenode.net irc.freenode.net) 00:43:19 --- join: michaelw (n=michaelw@lambda.foldr.org) joined #forth 00:43:26 --- join: astecp (n=aspect@210.8.131.1) joined #forth 00:44:35 --- join: Quiznos (i=1000@69-168-231-199.bflony.adelphia.net) joined #forth 00:44:47 re 00:46:21 --- join: Quartus (n=trailer@CPE0001023f6e4f-CM013349902843.cpe.net.cable.rogers.com) joined #forth 00:46:22 --- mode: ChanServ set +o Quartus 00:48:08 --- join: ccfg (n=ccfg@dsl-roigw1-fe8ade00-21.dhcp.inet.fi) joined #forth 00:48:09 --- join: fission (n=fission@69.60.114.33) joined #forth 00:48:11 --- join: uiuiuiu (i=ian@dslb-084-056-234-056.pools.arcor-ip.net) joined #forth 00:48:12 --- join: cods (n=cods@tuxee.net) joined #forth 00:48:15 wb 00:52:17 --- quit: Quiznos (sterling.freenode.net irc.freenode.net) 00:52:18 --- quit: k4jcw (sterling.freenode.net irc.freenode.net) 00:52:18 --- quit: ohub (sterling.freenode.net irc.freenode.net) 00:52:20 --- join: Quiznos (i=1000@69-168-231-199.bflony.adelphia.net) joined #forth 00:52:59 --- join: k4jcw (n=jcw@adsl-065-006-151-062.sip.asm.bellsouth.net) joined #forth 00:56:46 --- join: AI_coder (n=AI@ip-209-124-242-76.dynamic.eatel.net) joined #forth 00:57:22 Is there a word that converts an xt to a string or simply prints the name of the xt? 00:57:46 Not standard. 00:57:52 gforth? 00:57:54 what does xt point to? 00:58:00 the name of a word? 00:58:10 or xt >name> 00:58:13 oops 00:58:18 or xt >name 00:58:22 Example: ' + showxtname + ok 00:58:46 .id is the usual word that prints a word's name 00:58:58 duno if .id is ANS tho 00:58:59 ai_coder - yes, not standard. I wrote one recently that worked well with internal gforth words, and applied only to colon definitions. 00:59:45 Quartus what field does an xt point to? 00:59:54 ' words >name .name will display 'words' in gforth. 01:00:05 kk 01:00:30 Quartus: Thanks that works nicely. 01:00:43 An xt is a value that can be passed to execute to run the word. It has no definition in terms of pointing at a field. In a native-code system, it very likely points directly at the executable code. In other kinds of threading models, it may differ. The standard allows implementation freedom. 01:01:12 ok, so it would be the equivalent of the CFA 01:01:35 since that's the start of the body of a word 01:01:51 It may be anything at all that suits the implemenation. It could be an index into a bytecode table. It could be the header itself. It's up to the implementation. 01:02:00 sure 01:02:42 xts aren't necessarily unique; more than one word may have the same xt. 01:03:06 wouldnt that be the defining-word that is shared? 01:03:42 That question implies a specific implementation model. It actually has no meaning for a number of types of implementation. 01:03:52 Quartus: You're good, that was pretty impressive how you pulled that one off the top of your head, when I look through the manual it's there word for word, but it is so hard to remember all of the manual at once or do an intelligent query of the manual. 01:04:07 AI, glad it helped. 01:06:03 --- join: Cheery (n=Cheery@a81-197-19-23.elisa-laajakaista.fi) joined #forth 01:12:20 --- join: ohub (n=oherrala@sikw1.oulu.fi) joined #forth 01:39:01 --- join: nighty (n=nighty@sushi.rural-networks.com) joined #forth 01:46:59 --- quit: nighty (Read error: 104 (Connection reset by peer)) 01:48:36 Functional languages use pattern matching to select the appropriate form of a function, like f n = 0; f n = n * f(n - 1) is this style of programming possible in forth? 01:49:09 i'd guess at a higher-level bc that's one direction i'm heading 01:49:23 Joy does fp i think 01:49:27 a kind of fp 01:50:23 I just go crazy when I make if else conditional then statements in forth, I can't read them. 01:50:53 It's probably just me though. 01:50:58 then break to a newline after the keyword 01:51:04 with indenting 01:51:15 ie, 01:51:17 Yeah white space does help. 01:51:18 test if 01:51:24 \t truepart 01:51:29 else 01:51:31 \t falsepart 01:51:35 fi 01:51:43 fi? 01:51:50 oops, a bashism 01:52:06 economy of characters 01:52:28 Indeed, I would prefer a fi. 01:52:35 then alias it 01:52:38 How? 01:52:39 ' then is fi 01:52:45 or something that way 01:52:55 or, : fi compile then ; 01:52:57 or something that way 01:53:02 1 ALLOT ?Not building 01:53:02 1 ALLOT<-- 01:53:06 bloody hell 01:53:21 stop poking the forth 01:53:21 :) 01:53:25 heh 01:53:31 * fission pokes it inquisitively 01:53:51 you b0rkd it already 01:53:55 give it a rest :) 01:54:05 so I figured that I can stuff things in random memory locations with ! and grab it back with @, only I have no idea how to actually allocate memory heh 01:54:23 hehe 01:54:24 ranmdom pokage is not good 01:54:30 but it seems so right! 01:54:34 lol 01:54:47 allot 01:54:56 or allocate 01:55:17 do you know how many cells you need? 01:56:06 well allot just says ?Not building 01:56:18 Your machine may not be 1 au aligned, you may need to address at 4 au aligns. 01:56:31 how big are your cells? 01:56:32 4 ALLOT ?Not building 01:56:32 4 ALLOT<-- 01:56:32 What type of hardware are you on? 01:56:37 AI_coder: pdp-10 01:56:38 :P 01:56:38 bytes/cell? 01:56:47 uh 01:56:48 that's the twenex system? 01:56:51 36-bits per word 01:56:51 fission, at a guess it's sufficiently broken to allow allot only after create. 01:56:56 heh 01:56:59 Quartus: there's no create 01:57:12 oh, right. The equivalent from back then... let me check. 01:57:13 or if there is a create, it's not called create 01:57:19 build> iirc 01:57:28 it's a FIG forth, isnt it? 01:57:38 beats me heh 01:57:46 man i dont know where my fig books are at this hour 01:58:11 Does cells work? 01:58:16 1 cells allot 01:58:18 there's no cells 01:58:20 there are no cells in fig 01:58:34 try : foo ; 01:58:45 AI_coder have you found src for that binary yet? 01:58:51 I've figured out that SW,,AP seems to : FOO ; Ok 01:58:55 --- quit: segher (Read error: 60 (Operation timed out)) 01:58:56 oops heh 01:59:01 : FOO ; Ok 01:59:08 then foo bar 01:59:10 Quiznos: What src? 01:59:17 for the fig you're on 01:59:18 foo bar Ok 01:59:21 then foo . 01:59:23 holy crap 01:59:24 sorry 01:59:26 then bar . 01:59:31 --- join: segher (n=segher@dslb-084-056-159-011.pools.arcor-ip.net) joined #forth 01:59:31 bar .8992 Ok 01:59:51 So bar is now a word that returns the address of four ... bytes? words? of memory. 02:00:10 heh probably 4 36-bit words 02:00:18 or 4 18-bit half-words 02:00:29 a vector of four words 02:00:30 but who knows, really? 02:00:35 the [0] of that vector 02:00:37 Maybe. Anyway, that's the deal. The system you've got there refuses to allot unless it's against wow 02:00:55 so FOO is now a general-purpose allocator for a 4-bit word? 02:00:58 er no 02:01:03 a 4-something 02:01:05 we don't know what 02:01:06 If you do foo zot and then zot . you'll get a somewhat more incremented address. 02:01:23 foo zot Ok 02:01:23 zot .9000 Ok 02:01:37 So that's working. 02:01:46 my guess is that it's in half-words 02:01:58 so that was 4 36-bit words, or 8 halfwords 02:02:07 Experiment should reveal that. 02:02:41 right 02:02:51 you should really be applying your brain to learning & using a Forth that works. :) 02:02:58 ahahaha 02:03:01 but this is so cool! 02:03:07 Yeah, unless using old computers is uber. 02:03:13 Yes, but it's hurting me to remember this ancient and sad crap. 02:03:31 send me your FIG books plz kthx :P 02:03:55 umm 02:04:02 i cant 02:04:04 i love books 02:04:11 i have law books 02:04:19 but i'd like a little remuneration for them :) 02:04:19 ahaha 02:05:00 you can find a zipfile of fig and other docs online tho; mayhaps Taygeta has them 02:05:17 I will only pay good money for one book: "Finer Points in the Spacing & Arrangement of Type", by Geoffrey Dowding 02:05:20 i know taygeta has them 02:05:33 ok I'll check that out 02:05:39 this forth is hurting my face 02:05:49 stop kissing the crt 02:05:56 ahahahaha 02:05:58 * fission disengages 02:09:10 A traditional FIG Forth wouldn't care if you were doing a yeah it's the bastard sword of forth 02:10:10 Not what I meant. 02:10:14 ahahaha 02:10:40 but you have to admit that making it do anything useful at all without completely thrashing itself is fun 02:10:59 Not especially fun for me. I'd sooner be helping you with something useful and current. 02:11:29 tell ya what, if I figure out the return stack on this thing, I'll switch to pforth or gforth or something and make something useful 02:11:47 coward :) 02:11:48 lol 02:12:03 buck the trend, be original, use FIGforth 02:12:06 lol 02:12:15 ok....... : zap 3 5 >r . r> . ; 02:12:22 actually, ciforth is a figgy forth 02:12:42 there's no data at . 02:13:19 >r >R? 02:13:29 >r r> two different words 02:13:31 it doesn't know that one 02:13:38 r> R>? 02:13:41 or that one 02:13:56 oh. Ok, it's lost without a map. That's a FIG word, so the damaged author has left it out. 02:14:02 ahaha 02:14:14 Maybe send me the vlist in a msg. 02:14:14 <-,, SW,,AP ,,-> 02:14:17 ok 02:23:01 * fission yawns 02:23:15 damn I think it's officially too late for me to go to bed 02:29:41 --- quit: AI_coder (Client Quit) 03:06:44 heh wow this really is a 36-bit forth 03:06:55 cool 03:07:12 34359738367 2 + .-34359738367 Ok 03:07:24 signed too 03:07:26 yeah 03:07:29 try u. 03:07:31 check this out though: 03:07:51 34359738367 1+ . --,-+')-(-*( Ok 03:07:53 heh 03:07:59 there's no u. 03:08:16 is it an upcase forth? 03:08:28 you can define a u, 03:08:30 u. 03:08:36 upcase? 03:08:42 uppercase 03:08:42 it's case insensitive 03:08:45 k 03:08:50 I think it toupperizes everything though 03:08:53 yea 03:09:05 haha I just typed 'see .u' in bc 03:09:13 you need to browse taygeta.com now for fig code and docs 03:09:32 oh! 03:09:39 actually 03:09:45 this might work, one sec 03:09:51 ignore the system names when looking for fig files 03:10:02 like, dos, amiga, 68k, linux, etc 03:10:42 --- join: nighty (n=nighty@sushi.rural-networks.com) joined #forth 03:11:00 see (ud.) 03:11:00 ( 808DCD0 ) <# #S #> ; 03:11:09 ok 03:11:15 (that's on pforth) 03:11:18 k 03:11:26 but it doesn't work on this one heh 03:11:32 what does it do? 03:11:50 --- join: neceve (n=claudiu@unaffiliated/neceve) joined #forth 03:12:22 well on pforth it's: 03:12:26 see u. 03:12:26 ( 808DD70 ) 0 UD. ; 03:12:30 see ud. 03:12:30 ( 808DCE0 ) (UD.) TYPE SPACE ; 03:12:45 k 03:12:45 then it's that ud above 03:13:00 so I guess I'm supposed to put a 0 on the stack first? 03:13:13 i dont remember 03:13:38 try it 03:13:44 -1 Ok 03:13:44 0 <# #S #> . 0 Ok 03:13:46 it's forth 03:14:22 I don't even have a clue what <# or #S or #> does 03:14:39 <# peels one digit out of the tos 03:15:01 #s continues the conversion appending the resulting digit to the string being built 03:15:09 #> ends the convsion 03:15:37 tay-geta tay-geta tay-geta tay-geta 03:15:40 heh 03:16:06 hmm I dunno why I get 0 03:16:13 i think it's the remainder 03:16:16 1191182766 03:16:17 oh I see 03:17:13 hmmm I wonder where that number came from 03:58:32 Ok I give up 03:58:42 this thing can't even print unsigned numbers without a fuss 04:31:55 --- join: vatic (n=chatzill@pool-162-83-254-201.ny5030.east.verizon.net) joined #forth 04:41:46 --- join: PoppaVic (n=pete@0-1pool72-89.nas24.chicago4.il.us.da.qwest.net) joined #forth 04:51:57 fission just learn what the words do. unsigned printing is NOT a problem, it's just your lackage of knowing the words 04:52:29 hmm? 04:53:24 fission is tryin to u. and the fig he's on wont 04:53:46 fig-forth?? 04:53:48 It should 04:53:50 36bit cells 04:53:56 i know, that' what i'm tellin him 04:54:09 he doesnt know the meaning of the forth words 04:54:26 mind you, w/o the info-file for Gforth (and before that PFE), I'd have been horribly lost as well. 04:54:40 he's not using gforth 04:55:29 --- join: virl (n=virl@chello062178085149.1.12.vie.surfer.at) joined #forth 04:55:30 well, the point is the same: you need the info/man pages 04:55:38 true 04:55:49 ..I don't think _anyone_ uses "view fields" anymore, do they? 04:55:51 i encouraged him to scour taygeta for fig info 04:56:08 yep, w/o some docs - he's toast. 04:56:25 well, no, he can still spermint 04:56:33 but it will take longer 04:56:55 Depends on what he's trying to learn.. Certainly the FIG source is easier than some systems. 04:57:06 the whole i spose 04:57:09 ..trying to peruse gforth source makes you want to cry 04:57:28 yea, i wont even ftp gforth 04:57:46 but if i do, it might be to grab forth source from it 04:57:52 well, once it's installed, it's handy 04:58:03 perhaps 04:58:35 it's like a cross between a shell and bc, etc 04:58:45 k 04:59:25 I blip into it and use 'see' and bases and printing quite a bit.. Beats trying to find my calculator, it's manual and all the other noise. 04:59:56 true 05:00:03 makes better sense to me than bc or dc 05:00:27 well, bc CAN be handy, but I like doodling with forth 05:00:39 yea 05:01:01 and, of course, like C: if I do something the same a few times, I can just write a secondary. 05:01:22 yep 05:02:46 hmm.. 05:03:15 I'm glaring at my Klasses and trying to get my head together, so's I can beging to THINK about implementing. 05:04:23 I believe at some point I got a trifle disoriented, and atop that - in some cases I seem to have gotten confuzzled. 05:05:22 For example.. I can't seem to admit to myself that arrays should be "growable" 05:06:44 Seems to me that arrays are just that - and if they want "growth", then they need some other sort of container. 05:08:18 well, for growable vectors you're gonna need a vector ** type of object 05:08:22 or even *** 05:08:40 so that you have atleast one stable pointer to use 05:10:24 I'm still not 100% on if they _should_ grow. An "array" seems nice and simple to me - if they want a list or tree or stack or something, then that's what they should use. 05:10:41 what about a queue then? 05:10:59 same as the latter choices, of course 05:11:10 are you just filling out your kernel or going somewhere with grown vectors? 05:11:19 instance is an item; an array is a limited-set of item 05:11:38 just trying to coordinate basic "core" classes 05:12:23 k and k 05:12:29 aamof, I also been glaring at the idea of "array" - and it occurs to me that (in our objectified universe), I don't even need to force a class of elements. 05:12:58 k 05:12:59 ..as long as they are all "objects", they may all be shoved into the "array" 05:13:10 ..or, so it seems 05:13:14 they're all pointers ultimately 05:13:19 yeppers 05:13:38 ..and I do NOT plan to expose ptrs to uses. 05:13:40 users 05:13:46 k 05:13:59 but that's just a contrivance; the user will figer it out :) 05:14:11 dont use too much wool :) 05:14:51 if they want to dip into the interface, and write themselves some C or asm, fine. BUT, the idea of unleashing C's ptrs on the stacks and for users is just not thrill-making 05:15:15 I know they WILL be around, but call them XT's or some damned thing 05:15:38 i think you're tryin to hard 05:15:44 ..otherwise, some idjit will trying poking and proding areas they should avoid 05:15:49 heh 05:38:05 --- quit: vatic ("*poof*") 06:12:07 found an tarball of f83 ported to 68k; it's old and i now need to see if it can be ported 06:12:17 heh 06:12:18 to intel platform 06:12:44 Umm, I did F83 (notice case) for my DOS machine years ago 06:13:33 I'd expect the asm - anywhere but DOS or whatever - is the prime issue 06:14:47 Quiznos: How would to define/contrast [un]{ordered,sorted}? 06:15:12 http://www.forth.org/library/eforth_SOC/eforth_SOC_source/f83/f83v2-ms.lzh 06:15:18 Why are you reinventing the wheel? 06:15:27 F83 for x86 06:24:46 damn.. to/you 06:29:57 Google. It's almost like magic. 06:42:44 ahhh 06:50:37 PoppaVic there is order (rank, sequence, a stepping) to everything, even if the the order is unrecognised or unknown. c++ ``bags'' are seen as unordered but they could be temporally-sorted as to when the elements (groceries) were placed in the bag 06:50:57 that's how i see it 06:51:15 yep... That was the feeling I got from a short search of google and ##C debate 06:51:21 ok 06:51:30 i'll bbl. 06:51:34 sure 06:52:07 on the flip side, there is no real "chaos" because that requires confusion and conputers dont do that :) 06:52:10 lol 06:52:32 no, people 06:52:38 true chaos exists only within a bureaucracy. 06:54:44 brb, guys - need the horn... 06:54:46 --- quit: PoppaVic ("Pulls the pin...") 06:59:50 --- join: madwork (n=foo@derby.metrics.com) joined #forth 07:02:05 Quiznos: I can't tell whether you were being deliberately rude or just lack tact 07:09:36 fission - 'deliberately direct'. Best not to speculate about tact. 07:28:18 --- join: PoppaVic (n=pete@0-1pool64-131.nas22.chicago4.il.us.da.qwest.net) joined #forth 07:53:46 --- join: I440r (n=mark4@63.163.143.188) joined #forth 08:02:27 ayrnieu: haha clever -- I'll keep that in mind 08:03:01 Quiznos: just to clarify: yes, I know that I need to know what the words mean -- if this forth had "see", that would not be a problem 08:03:21 Quiznos: I have poked around taygeta but didn't really see anything useful 08:04:04 Quiznos: I can't see words to manipulate a return stack, and <# #S #> seems not to know about double-length numbers 08:04:27 but, as I said, I'm giving up on a forth without documentation and missing chunks of words that I'm used to and fond of 08:05:40 FIG forths are complete. They predate most everything except C and lisp 08:06:00 I'm not saying that FIG forths aren't complete 08:06:11 I'm saying that this particular implementation appears to be incomplete 08:06:14 meanwhile, you can't live w/p source and docs - even if the docs *are* the source 08:06:31 yeah there's no docs, no source 08:06:44 there is no "incomplete" to FIG - there are just extensions you do not recognize 08:06:47 so I'm basically typing words to see what they do and inferring what its behaviour is 08:06:57 fission, isforth has see :) 08:07:15 so does Gforth - and prolly Pfe 08:07:28 PoppaVic: so there's no basic set of words that a FIG forth should implement? 08:07:36 I'll probably use pforth 08:07:37 neither of which counts as a real forth to me 08:07:42 both are written in c 08:07:54 fission: depends on yer context: ANS specifies some, FIG does as well 08:07:54 I440r: heh so what forth do you recommend? 08:08:07 PoppaVic: where can I see the FIG spec? 08:08:27 I440r: yeah, asm is god - until the asm is not apropos for the cpu or os 08:08:27 isforth.clss.net - a work in progress 08:08:35 but then im biased heh 08:08:39 there IS no "FIG 'spec'" 08:08:57 ok so what I'm hearing is that there is no requirement for any word in a FIG forth 08:08:58 there is FIG - and there are others 08:09:08 PoppaVic, and portable code is a myth and asm is damned SIMPLE to port 08:09:14 so wheres the problem 08:09:21 forth is "the world most unsupported language" 08:09:43 I440r: assumptions are the mother of fuckups 08:09:53 the whole thing seems really silly anyhow 08:10:14 fission: sure, depending on yer assumptions and viewpoints 08:10:17 --- mode: ChanServ set +o I440r 08:10:24 I say "this forth is incomplete" and you (without ever having used it, I assume) say "that couldn't possibly be" 08:10:34 so when I ask how that is, you say there is no spec? 08:10:35 ... 08:10:36 what? 08:10:49 "incomplete" to one clone is "flexible" to another - so what? 08:11:41 basically you're talking a lot but not saying much 08:11:42 that's what 08:12:03 FIG was designed for 8080, 6502 and 6510, and then Z80's - SO WHAT? Ohhh... yer freaked on the OS 08:12:38 --- quit: madwork ("?OUT OF DATA ERROR") 08:12:46 forth-purists indicate they feel every forth is unique and nonportable. 08:13:21 --- join: madwork (n=foo@derby.metrics.com) joined #forth 08:13:31 Seems silly to me, as there are basic assumptions everywhere. 08:13:43 PoppaVic: "yer freaked on the OS" ? 08:14:11 fission: every OS and platform offers "challenges" - and then has to mesh 08:16:09 I get folks pissed in ##C and #forth and even #asm - because there are basic expectations, then solutions or assumptions - both. And no one wants to admit they are above the OS and the kernel and bios and interpreting - one way or another. 08:16:56 ..the only OTHER alternative is: you just became the kernel and OS 08:17:24 ..and they STILL have to talk-with or learn-from the BIOS 08:17:45 ..let alone the cpu 08:18:39 I've long suspected every language could be forthish-based. But, the proofs are hard to find. 08:19:09 linus t has a friend thats developing a dual stack c compiler 08:19:26 linus himself has done forth coding but he doesnt like forth because it has no data types 08:19:43 which are a crutch for poor programmers anyway (chuck moore) 08:21:27 --- quit: I440r ("brb - restarting x") 08:21:43 chuck started getting whachy post FIG.. I don't much care 08:22:27 --- join: I440r (n=mark4@63.163.143.188) joined #forth 08:22:41 Chuck esposes Color-Forth - which made me laugh no end 08:23:48 btw did anyone here notice the new joke at isforth.clss.net ??? :) 08:23:54 I don't care what you write in: structures, unions; methods and funcs and procedures are still your "issue" 08:24:49 --- mode: ChanServ set +o I440r 08:24:59 I get a laugh with most ASM, because - by themselves - they still have the same issues even libc and vars/structs offer 08:25:57 so, (once again), we need a generic macro-assembler, plus preprocessing - and Forths offer mode and state anyway 08:26:34 PoppaVic, not everyone thinks you need 500 gigs of library glue code to write programs 08:26:45 I440r: that's nice. 08:26:50 chuck moore doesnt give a damn about writing code that is portable 08:26:58 meanwhile, that code exists 08:27:18 I440r: yes, and thus: I'm not much interested 08:27:34 we know 08:27:38 ..I watch for the "issues" - and the rest is dross 08:28:18 PoppaVic, in forth, the overhead for writing something from scratch is NEGLIGABLE !!!!!!! 08:28:38 I440r: I would hope so. I disagre with a lot of folks, and I don't much care whome disagrees - the WHY is more amusing 08:29:09 PoppaVic, sometimes us lesser mortals have difficulty understanding you lol 08:29:11 I440r: yer silly... Yer thinking of interp, not compile or object or share. 08:29:40 I440r: I keep looking at does and what and whom and then WHY 08:29:59 ? 08:30:04 yeah 08:30:07 say again in english ? 08:30:09 I thought so 08:30:40