hate these ads?, log in or register to hide them
Page 119 of 119 FirstFirst ... 1969109116117118119
Results 2,361 to 2,380 of 2380

Thread: <xml><thread isTerrible="true" isUnderstandable="false" /></xml> - Coding help thread

  1. #2361
    Donor halbarad's Avatar
    Join Date
    April 9, 2011
    Posts
    4,779
    Just remember to use lots of regex.

  2. #2362
    Movember 2012 I Legionnaire's Avatar
    Join Date
    April 9, 2011
    Posts
    1,573
    This... thing was used to count up the votes in the 2014 Polish election.

    https://github.com/wybory2014/Kalkul...intProtocol.cs

  3. #2363

    Join Date
    April 13, 2011
    Posts
    5,324
    Quote Originally Posted by I Legionnaire View Post
    This... thing was used to count up the votes in the 2014 Polish election.

    https://github.com/wybory2014/Kalkul...intProtocol.cs
    holy fucking shit

  4. #2364
    root's Avatar
    Join Date
    April 26, 2011
    Location
    The Camel Empire
    Posts
    2,887
    Poland is a country in which the fate of thousands of commissioners lies on the shoulders of a beginner programmer.
    :/
    The Rapier is my love boat
    ~lowsec smallscale pvp 'n stuff~

  5. #2365
    NoirAvlaa's Avatar
    Join Date
    April 12, 2011
    Location
    Liverpool, laaaa
    Posts
    4,196
    Quote Originally Posted by elmicker View Post
    Quote Originally Posted by I Legionnaire View Post
    This... thing was used to count up the votes in the 2014 Polish election.

    https://github.com/wybory2014/Kalkul...intProtocol.cs
    holy fucking shit

  6. #2366
    dominus's Avatar
    Join Date
    October 5, 2012
    Posts
    1,898
    Quote Originally Posted by elmicker View Post
    Quote Originally Posted by I Legionnaire View Post
    This... thing was used to count up the votes in the 2014 Polish election.

    https://github.com/wybory2014/Kalkul...intProtocol.cs
    holy fucking shit
    A 3000 line class full of properly written, formatted and commented code, what could go wrong?

    You all are just jealous /s

  7. #2367
    Frug's Avatar
    Join Date
    April 9, 2011
    Location
    Canada
    Posts
    12,932
    Quote Originally Posted by NoirAvlaa View Post
    Quote Originally Posted by elmicker View Post
    Quote Originally Posted by I Legionnaire View Post
    This... thing was used to count up the votes in the 2014 Polish election.

    https://github.com/wybory2014/Kalkul...intProtocol.cs
    holy fucking shit

    Quote Originally Posted by Loire
    I'm too stupid to say anything that deserves being in your magnificent signature.

  8. #2368
    Daneel Trevize's Avatar
    Join Date
    April 10, 2011
    Location
    T L A
    Posts
    11,986
    Modern JavaScript Explained For Dinosaurs

    [IMG]https://cdn-images-1.medium.com/max/1600/1*H8PH-HaV43gZyBJz0mJHxA.png[/IMG]

    Edit: lol FHC/vB can't handle that URL for an image.
    Quote Originally Posted by QuackBot View Post
    Idk about that, and i'm fucking stupid.

  9. #2369
    Donor Sponk's Avatar
    Join Date
    April 10, 2011
    Location
    AU TZ
    Posts
    11,307
    yeah, not a bad article.

    I guess it's my turn.

    https://blog.acolyer.org/2017/10/19/...ctural-impact/
    Contract stuff to Seraphina Amaranth.

    "You give me the awful impression - I hate to have to say - of someone who hasn't read any of the arguments against your position. Ever."


  10. #2370
    Daneel Trevize's Avatar
    Join Date
    April 10, 2011
    Location
    T L A
    Posts
    11,986
    Contrast this with server-based designs, where you pay per (long-running) instance. In such a world it is hard to justify having a dedicated service instance for infrequent but important tasks (let alone two if you have a primary and fail-over). You have a strong economic incentive to bundle responsibility for many tasks into the same instance.
    Virtual Machines m8s, you might have heard of them...
    This means that it is perfectly acceptable, even expected, to allow client applications to directly access resources traditionally considered ‘back-end’. AWS provides several distributed authentication and authorization mechanisms to support those connections. This is a major change from the traditional client/server model, where direct access from clients to back-end resources, such as storage, is a major security risk.
    Maybe if we don't call it the server, we won't have to do validation... Can only end well.
    Last edited by Daneel Trevize; October 19 2017 at 02:50:58 PM.
    Quote Originally Posted by QuackBot View Post
    Idk about that, and i'm fucking stupid.

  11. #2371
    Donor erichkknaar's Avatar
    Join Date
    April 9, 2011
    Posts
    8,670
    Quote Originally Posted by Daneel Trevize View Post
    Contrast this with server-based designs, where you pay per (long-running) instance. In such a world it is hard to justify having a dedicated service instance for infrequent but important tasks (let alone two if you have a primary and fail-over). You have a strong economic incentive to bundle responsibility for many tasks into the same instance.
    Virtual Machines m8s, you might have heard of them...
    You need to spend some time understanding how it works, before poo-pooing it m8.

    VMs. Hmm. So to run a method that returns a web response, lets say a json doc, in an HTTP response. I need a kernel, a set of libraries providing all the functionality I don't want to create myself, some business logic to find and send back the right json doc (lets assume you are using a db with a rest interface, so nothing with massive connection startup cost). finally, you need an HTTP request handler of some kind. Now, in any real application, you have a load balancer in front of this, which is is handling the internet and offloading requests to multiple back ends. It used to be we had to deploy a separate machine to do each of the backends, with liek maybe tomcat, or somethign on it, and it was super complicated to predict your traffic and if you got it wrong, you had a ton of hardware spare and you wasted a bunch of money, OR, you got it the fuck wrong, and you look like idiots with broken shit, because the load crushed you. Worse was if you has very spiky traffic, because then you have a massive capacity that isn't used a lot of the time. VMs allowed us to consolidate this to a few bigger hosts, and thus the electricity bill went down.

    In addition, Your VMs are probably waiting for traffic when there is none. Depending on your scalability and redundancy requirements, you may actually need to split your application up and ensure a minimum number are available as well, now you have to have like, 10 VMs at minimum, on all the time. That gets expensive.

    So, Docker. Ok, VMs exploded, just like hardware used to, and running 5k reserved instances on AWS is stupidly expensive, so we need to consolidate, AGAIN, and those clever chaps at Joyent built this wonderful Solaris-derived OS called smartOS, which leveraged zones in a fancy newfangled workflow. Now, not everyone can afford the kind of admins you need to actually run Solaris in production, so it kind of has a built cap on the user numbers, and also you have to develop on Solaris, which, isn't bad, but it isn't Linux, so some dudes figured out that all the cgroups stuff Google was putting into the Linux kernel could basically do the same thing and joyent were using the trick you can do with gzipping ZFS file systems and the work AUFS was doing to replicate some of the block differential mechanisms Sun and Netapp got into a slapfight over, and suddenly Docker.

    Now we can consolidate to some interesting ways onto a smaller core cluster, and using some schedulers, (mesos, fleetd, etc), just start to treat our cluster as a pool of compute, memory and storage and kind of divvy it up in a much more dynamic way, almost like what high end HPC has been doing with very expensive interconnect technology and things like NUMA. Well, our understanding of distributed systems has come on too, and we have some new ways to think about architecture, and the next logical step is essentially building a system where we can treat this cluster like an execution environment that allows us to spawn processes in response to events, sort of how POSIX works with signals.

    So a bunch of work later we get serverless, and you are literally now just responsible for the handler portion of the code. You no longer have to give a shit about any of the rest of the run time. All you know is that the JVM or the Python interpreter or the Node interpreter or whatever will be there with there when the event that you want to call the function in response to, fires. Whether that's an HTTP event, or a pubsub message, or anything else, its just that most of the runtime, the load balancer, etc, is no longer something you have to spend resources on maintaining, In addition, it's all extremely easy to automate, with most of the expensive "configure" steps no longer being necessary. Plus, you pay for your resources by the microsecond. \

    It's 2 generations beyond VMs.

    Quote Originally Posted by Daneel Trevize View Post
    This means that it is perfectly acceptable, even expected, to allow client applications to directly access resources traditionally considered ‘back-end’. AWS provides several distributed authentication and authorization mechanisms to support those connections. This is a major change from the traditional client/server model, where direct access from clients to back-end resources, such as storage, is a major security risk.
    Maybe if we don't call it the server, we won't have to do validation... Can only end well.
    If you have a shit development process, nothing will help you. Instead of validation problems, you simply won't have developers updating 4 year old tomcat instances on out of date ubuntu VMs.
    meh

  12. #2372
    Daneel Trevize's Avatar
    Join Date
    April 10, 2011
    Location
    T L A
    Posts
    11,986
    The abstracting & outsourcing of the underlying OS & app server layers sounds smart, but it's assuming both that the provider is doing things perfectly (patching, scaling to demand), and isn't charging you more to cover having burst potential than you would pay yourself for a more traditional setup. But of course they are. For some this will will be a net gain in less IT (staff) costs, for some it's simply not an option unless the lower levels are also in-house, and it becomes just a different way to delineate between app devs and tech support.
    It'll be fun to play with when the hardware to run the whole setup trickles down to be ubiquitous (and it's FOSS), just as VM clusters let you fake a whole datacentre in a home PC setup.
    Quote Originally Posted by QuackBot View Post
    Idk about that, and i'm fucking stupid.

  13. #2373
    Donor erichkknaar's Avatar
    Join Date
    April 9, 2011
    Posts
    8,670
    Quote Originally Posted by Daneel Trevize View Post
    The abstracting & outsourcing of the underlying OS & app server layers sounds smart, but it's assuming both that the provider is doing things perfectly (patching, scaling to demand), and isn't charging you more to cover having burst potential than you would pay yourself for a more traditional setup. But of course they are. For some this will will be a net gain in less IT (staff) costs, for some it's simply not an option unless the lower levels are also in-house, and it becomes just a different way to delineate between app devs and tech support.
    It'll be fun to play with when the hardware to run the whole setup trickles down to be ubiquitous (and it's FOSS), just as VM clusters let you fake a whole datacentre in a home PC setup.
    http://fission.io/

    On the bolded part. Google has 700 employees with the word security in their title. AWS has a smaller, but still significant number. How many does your company have?
    Last edited by erichkknaar; October 19 2017 at 05:28:37 PM.
    meh

  14. #2374

    Join Date
    April 13, 2011
    Posts
    5,324
    Quote Originally Posted by Daneel Trevize View Post
    The abstracting & outsourcing of the underlying OS & app server layers sounds smart, but it's assuming both that the provider is doing things perfectly (patching, scaling to demand), and isn't charging you more to cover having burst potential than you would pay yourself for a more traditional setup. But of course they are. For some this will will be a net gain in less IT (staff) costs, for some it's simply not an option unless the lower levels are also in-house, and it becomes just a different way to delineate between app devs and tech support.
    It'll be fun to play with when the hardware to run the whole setup trickles down to be ubiquitous (and it's FOSS), just as VM clusters let you fake a whole datacentre in a home PC setup.
    Alright, grandad

  15. #2375
    Daneel Trevize's Avatar
    Join Date
    April 10, 2011
    Location
    T L A
    Posts
    11,986
    Quote Originally Posted by erichkknaar View Post
    How many does your company have?
    When you have to have physical redundant DCs 100km apart, and keep all records (read: sub-hourly backups) for 7+ years, and everything has international risk assessment & change management, you don't leap on every new abstraction fad.
    Again, I said I'm sure it's immediately great for some. But not everyone can just slice off another part of the app and hand it off to the cloud.
    When investment firms are outsourcing to you, it's another ball game.

    Yes, it'd be nice to get away from some parts being on a run-thrice-daily uber batch of COBOL... but that shit underpins Western society m8.
    Last edited by Daneel Trevize; October 19 2017 at 08:05:01 PM.
    Quote Originally Posted by QuackBot View Post
    Idk about that, and i'm fucking stupid.

  16. #2376
    Movember 2012 I Legionnaire's Avatar
    Join Date
    April 9, 2011
    Posts
    1,573
    We call those availability zones in cloud-land.

  17. #2377

    Join Date
    April 13, 2011
    Posts
    5,324
    Quote Originally Posted by I Legionnaire View Post
    We call those availability zones in cloud-land.
    Regions. AZs are best thought of as rows in an end-of-row setup. Treating an AZ as a DC is asking for bother.

    This is one of the things Azure does much, much better than AWS. Full two dimensional availability settings for eeeerything.

  18. #2378
    Donor Aea's Avatar
    Join Date
    April 13, 2011
    Location
    Colorado
    Posts
    13,723
    Fuck Azure


    Sent from my iPhone using Tapatalk

  19. #2379
    Frug's Avatar
    Join Date
    April 9, 2011
    Location
    Canada
    Posts
    12,932
    Quote Originally Posted by Aea View Post
    Fuck Azure


    Sent from my iPhone using Tapatalk
    Fuck aws.

    Quote Originally Posted by Loire
    I'm too stupid to say anything that deserves being in your magnificent signature.

  20. #2380
    Donor halbarad's Avatar
    Join Date
    April 9, 2011
    Posts
    4,779
    Quote Originally Posted by Frug View Post
    Quote Originally Posted by Aea View Post
    Fuck Azure


    Sent from my iPhone using Tapatalk
    Fuck aws.
    Fuck Google cloud? Or Oracle cloud, or Vodafone cloud or any cloud? (am I doing it right?)

    I don't mind Azure but I spend most of my time dealing with IaaS sadly so don't get to play around with the PaaS features or the other stuff there.

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •