|
Cybernetic Vermin posted:a big part is really that sun had excellent engineers and very little business sense. really solid investment into something that predictably got them pretty much nothing sun made a non-trivial amount of money off of java, hundreds of millions a year. the "problem" is that other java vendors were making billions a year. sun's java revenue mainly came from penny-ante mobile poo poo. sun's competitors licensed java for use in highly lucrative enterprise "middleware." ibm and bea were each doing a billion plus a year selling j2ee implementations sun tried again and again to get its beak wet in the middleware market but it never panned out for them. sun owned two j2ee implementations i know of ( ONE/iplanet/jsas + glassfish) and neither one was ever a big business
|
#
?
Oct 25, 2014 17:19
|
|
|
|
| # ? Jun 11, 2024 18:35 |
|
|
the middleware business has more or less collapsed now i would be very interested to learn how much oracle makes off of their stable of j2ee implementations (oc4j, one/iplanet/jsas, glassfish, bea/tuxedo) same for redhat/jboss and ibm/websphere
|
|
#
?
Oct 25, 2014 17:21
|
|
|
More interesting TIBCO going private, but it's been an overdue end to the ridiculous $500k/crappy product premium.
|
|
#
?
Oct 25, 2014 17:38
|
|
|
MrMoo posted:More interesting TIBCO going private, but it's been an overdue end to the ridiculous $500k/crappy product premium. i have never worked with tibco software, but i always assumed it was just the 1990s version of CA make one product successful enough to jumpstart acquisitions, then just buy poo poo and put customers under thumbscrews until you have enough money for the next deal
|
|
#
?
Oct 25, 2014 18:32
|
|
|
there are so many things so much worse than tibco stuff though if anything i rather respect them for doing enterprise software which does not go quite all out to be obscure and artificially impressive-looking, if anything their stuff gets hated on because it is expensive and overtly pretty simplistic, but simplistic is a pretty drat good feature for enterprise software
|
|
#
?
Oct 25, 2014 18:48
|
|
|
Generally their software has been nice, the rest of the world is slowly picking up the event bus and service mentality as it is ridiculously nice to upgrade and replace components individually. I always love comparing the TIBCO based market data access with anything else, it is hilarious: TIBCO way: code:code:
|
|
#
?
Oct 25, 2014 19:18
|
|
|
reuters are incredibly incompetent and bloomberg thinks it is eternally 1992 thouh, so tibco has a real easy fight there mostly what i like about rv is that it is prosaic enough to be really pleasant to just randomly interact with. have a tiny little c program to query your services and explore as you please. very sensible (quirky but easy to work with) data model too tibco businessevents was a surprisingly ok piece of software too, though they seem to be pushing streambase instead now
|
|
#
?
Oct 25, 2014 19:22
|
|
|
multicast stormmmmm
|
|
#
?
Oct 25, 2014 19:33
|
|
|
Bumping this thread to remind everyone that Microchip is still a massive piece of poo poo
|
|
#
?
Oct 25, 2014 20:06
|
|
|
Mr Dog posted:Bumping this thread to remind everyone that Microchip is still a massive piece of poo poo Sounds like someone has a microchip on their shoulder
|
|
#
?
Oct 25, 2014 20:21
|
|
|
Wheany posted:take a look at http://www.greenfoot.org/door seriously look at greenfoot, it uses an actual programming language that people use irl (java) and it's all visual and interactive. https://www.youtube.com/watch?v=NcGe141R2yA&t=54s
|
|
#
?
Oct 25, 2014 20:25
|
|
|
90% of the kids that start with java are poo poo java programmers. The ones that start with a good teaching language come back to java and they're better at it after a year.
|
|
#
?
Oct 25, 2014 20:30
|
|
|
of course kids are poo poo programmers, they're kids.
|
|
#
?
Oct 25, 2014 20:43
|
|
|
"gently caress you kid, that is the worst candle holder i've ever seen, what do they even teach you at that woodcarving class!"
|
|
#
?
Oct 25, 2014 20:45
|
|
|
i assume he meant up through undergrad. i think every undergrad should have a programming intensive course where the focus is just pooping out as much code as possible and then having peers and educators review it. of course this requires motivated students and a lot of mentorship so would never actually happen in uni :/
|
|
#
?
Oct 25, 2014 20:45
|
|
|
Programming should unironically be taught the same way as creative writing. Bad programmers are bad because they can't structure code worth poo poo, and structuring is there for the benefit of other humans, not the computer, so there isn't really a scientific right or wrong way to structure a program.
|
|
#
?
Oct 25, 2014 20:52
|
|
|
Mr Dog posted:Programming should unironically be taught the same way as creative writing. Bad programmers are bad because they can't structure code worth poo poo, and structuring is there for the benefit of other humans, not the computer, so there isn't really a scientific right or wrong way to structure a program. yes there is. with tabs.
|
|
#
?
Oct 25, 2014 21:06
|
|
|
Cybernetic Vermin posted:tibco businessevents was a surprisingly ok piece of software too, though they seem to be pushing streambase instead now well, they bought StreamBase, they just need serious development effort to actually make it fault-tolerant and scalable: literally everyone today runs it as a single instance.
|
|
#
?
Oct 25, 2014 21:43
|
|
|
the point is that learning programming is not about learning a particular programming language
|
|
#
?
Oct 25, 2014 21:56
|
|
|
yes, but learning a programming language is better than learning no programming language.
|
|
#
?
Oct 25, 2014 23:22
|
|
|
fleshweasel posted:the point is that learning programming is not about learning a particular programming language you have to start somewhere, and to function in a classroom environment you gotta have everyone start with the same language creative writing is not limited to a single language, but you don't have composition 101 courses letting people go wild in german and french
|
|
#
?
Oct 25, 2014 23:30
|
|
|
Shaggar posted:way to recommend a bunch of p-langs so they start off down the wrong path. is ti basic considered p lang now. because thats where i started and im a poo poo webdev w/ no formal training
|
|
#
?
Oct 26, 2014 00:55
|
|
|
Mr Dog posted:Bumping this thread to remind everyone that Microchip is still a massive piece of poo poo agreed, mr. dog.
|
|
#
?
Oct 26, 2014 02:06
|
|
|
Blotto Skorzany posted:the universe of discourse in the subject of computer programming consists of quoting blithe one-liners first uttered forty years ago your posts considered harmful
|
|
#
?
Oct 26, 2014 02:16
|
|
|
It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration.
|
|
#
?
Oct 26, 2014 02:18
|
|
|
dijkstra was the dawkins of computer science
|
|
#
?
Oct 26, 2014 03:29
|
|
|
Subjunctive posted:your posts considered harmful much lke your posting! lol!
|
|
#
?
Oct 26, 2014 04:37
|
|
|
Wheany posted:seriously look at greenfoot, it uses an actual programming language that people use irl (java) and it's all visual and interactive. k I'll take a look at it and add it to the list. all I'm looking for is some stuff for high school kids to play around with to see if they even like programming at all. no reason to wait til you're halfway through a bachelors degree to realize that you hate computers fwiw the only programming I did before college was with a java book. all I remember is that I made a stupid little gui app that would calculate my target split times for track meets that was super simple but it felt good to make it
|
|
#
?
Oct 26, 2014 06:20
|
|
|
teach them hoon
|
|
#
?
Oct 26, 2014 06:32
|
|
|
suffix posted:It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration. but i got my start with both ti and q basic 
|
|
#
?
Oct 26, 2014 16:05
|
|
|
pithy one liners considered harmful
|
|
#
?
Oct 26, 2014 16:21
|
|
|
tef posted:tbh if the person wants to get minecraft modding, this might be an easier way to do it. although groovy is a bit rubbyesque for my taste and ends up in dsl hell you can write nearly pure java in groovy which is how I would encourage its use.
|
|
#
?
Oct 26, 2014 18:33
|
|
|
So lately I've been working on some specific cases of TCP systems being annoying as gently caress, specifically detecting a closed connection while sending onto a socket when the remote end has closed the connection but the local one isn't reading from the socket. It turns that depending on timing and whether data is waiting in the local buffer or not, some stacks will propagate a closed connection, some won't, some will after a time out, and sometimes you'll be stuck in a half-closed connection. Some optimizations seem to take place when you're talking on the loopback interface, some when doing things locally, and some behaviours only between remote hosts, but I'm not exactly too sure. Most of this is guest work because I've been able to reliably repeat some behaviours with specific set ups that won't work with others (a classical one is OSX localhost vs. a Travis CI node), but I'm wondering if anyone has good sources on that kind of stuff or differences. I'm not having the greatest time figuring them out by hand as they happen, and most resources approach TCP from a protocol perspective, not from its implementations first.
|
|
#
?
Oct 26, 2014 19:26
|
|
|
Shaggar posted:you can write nearly pure java in groovy which is how I would encourage its use. shaggar was right tef posted:tbh if the person wants to get minecraft modding, this might be an easier way to do it. although groovy is a bit rubbyesque for my taste and ends up in dsl hell gradle starts you off in dsl hell from minute one. gradle is often folks' first contact with groovy gradle is terrible, but groovy has a lot of potential
|
|
#
?
Oct 26, 2014 19:40
|
|
|
MononcQc posted:Most of this is guest work because I've been able to reliably repeat some behaviours with specific set ups that won't work with others (a classical one is OSX localhost vs. a Travis CI node), but I'm wondering if anyone has good sources on that kind of stuff or differences step one: minimize unnecessary differences. if you are deploying to linux, why on earth are you trying to do development work on osx localhost? you just spent serious time chasing down a problem that would never have existed if you did your development work on your target platform
|
|
#
?
Oct 26, 2014 19:48
|
|
|
Notorious b.s.d. posted:step one: minimize unnecessary differences. if you are deploying to linux, why on earth are you trying to do development work on osx localhost? Some of the problems happen across stacks, some happen within a single stack depending on what remote host you're talking to (what their stack is, which we don't necessarily control), and which interface you're listening to. At this point, short of having a list of things that may go wrong, I'd rather find about them as early as possible to be able to tackle them ASAP.
|
|
#
?
Oct 26, 2014 19:59
|
|
|
MononcQc posted:So lately I've been working on some specific cases of TCP systems being annoying as gently caress, specifically detecting a closed connection while sending onto a socket when the remote end has closed the connection but the local one isn't reading from the socket. You can't do it portably. We have Linux-specific code that listens for rdhup and in our case we treat it like an interrupted connection because you're not supposed to half-close the connection when talking to our server. That is important, because that lets us be able to interrupt the query we're evaluating without having to speculatively read from the connection and buffer all the information we get. I don't know if this handles every bad situation, but this is the best we could do. I think we also have heartbeats because there's other ways for connections to just up and disappear, I can't remember. We just ignore the problem if the server's running on OS X. If we used a fancier protocol on top of TCP then we could always read() until either the client's sent us too much data for us to buffer and we disconnect it or we see an EOF or EAGAIN. We'd have to be willing to negotiate with the client how much data it can send to us without asking permission for more, or in what circumstances it can send us more, or something, so that we can safely exhaust read(). Also with heartbeats. Edit: In the case of server-to-server connections, we have heartbeats, but not server-to-client. sarehu fucked around with this message at 22:44 on Oct 26, 2014 |
|
#
?
Oct 26, 2014 22:36
|
|
|
whoa i just found out you can use type hinting with arrays in php. not only that apparently it's been a thing since php 5.1??
|
|
#
?
Oct 26, 2014 22:53
|
|
|
Tiny Bug Child posted:whoa i just found out you can use type hinting with arrays in php. not only that apparently it's been a thing since php 5.1?? there's the foo(array $butts) thing, and phpdoc's @param Butts[] $butts, iirc which gives you hinting in fancy php ides
|
|
#
?
Oct 26, 2014 23:05
|
|
|
|
| # ? Jun 11, 2024 18:35 |
|
|
Tiny Bug Child posted:whoa i just found out you can use type hinting with arrays in php. not only that apparently it's been a thing since php 5.1?? types are for useless academics tbc, php always does the right thing with the contents of your arrays anyway so why would you need to janitor your array contents
|
|
#
?
Oct 26, 2014 23:27
|
|