|
Direct memory access is trivial in C#.
This space for rent
|
|
|
|
|
Thanks Pete. But on what context this is said ?
Is this about accessing unsafe* memory? Can this be really called "Direct Memory Access"? I'm not sure.
The only DMA I've heard of.. [^] 
Or it could possibly mean accessing cross-process memory, like in Kernel Mode programming.
Full Reset
|
|
|
|
|
Nathan Minier wrote: What I won't do is badmouth systems that do it better because their developers have the time or skill set to do better by using low-level code. That's just arrogant and preposterous.
They certainly 'claim' that it is better. Lots of people 'claim' they are doing it well. Every company I have ever worked at 'claimed' that their developers were above average. I have been in and heard of a large number of weird and odd interview processes that 'claim' will lead to better hires.
In my experience though developers have a very hard time even measuring performance of their own code. And that is something that can be objectively measured. Sometimes.
Kudos to the ones that manage to get it right. But I am certain that it more due to luck than actual thought. And most definitely due to building something on the ashes of the prior failures.
|
|
|
|
|
Nathan Minier wrote: You belittle the work of thousands of engineers based on a completely false premise and a lack of understanding, and I'm rude?
I did not belittle them. I wished them to do better.
If they were doing everything so brilliant & clear, why should they embrace Linux & open stack now?
They missed the train. They are just waking up and catching up.
I don't see anything wrong in it.
Been working with MS tools, frameworks, even with their teams directly. I do like them all.
I wish them to stay ahead.
Full Reset
|
|
|
|
|
Eytukan wrote: Micro-services, mainly after Node took over, YES, it's changing the landscape of a Web-server "module" & distributed computing.
Where "it" is what exactly? Micro-services? Micro-services are a really, really good way to end up in a world of hurt down the road. Not sure if most developers realize that but I have seen bad implementations of that and I am hopeful that others recognize that the required architecture (not technology) required to get it right slow them down or stop them before they attempt it.
|
|
|
|
|
I was actually adding this point in my previous reply. But I felt, I shouldnt speak more about a tech that I haven't worked on in detail. I'm just starting to use Node/Microservices. But there's precisely the problems popping up, making us think about how we need to share data across instances, how to sync them in real time & all that. I also do see, for every problem, there's a way and teams are getting to go past it.
Micro-services , for sure is over-marketed. But it definitely has a considerable goodness too.
Full Reset
|
|
|
|
|
Just noting that I didn't see anything in your response about how the architecture is being handled.
If a company does in fact have sufficient traffic or at least a reasonable (actual) expectation that they will have it in the future then if they fail at the architecture for the microservices then they will end up with serious problems.
The architecture must start with a very solid and detailed understanding of what the actual (not hypothetical) load of the business will be with real (again not hypothetical) load dynamics will be.
If that doesn't happen then 5 years down the road what the company ends up with is what started as an amusement for the developers (no business value) and now a maintenance problem on top of all the other maintenance problems that usually accumulate with legacy code.
|
|
|
|
|
Nathan Minier wrote: You're also confusing a runtime with a software module, which makes me wonder how anyone on this site (outside of QA) can be so uniformed about basic computer operation.
Perhaps because some of us actually understand how complex the industry has actually gotten.
Someone that has worked in nodejs for years focusing on human factors engineering is unlikely to need to know in detail that nodejs actually does use threads much less how it does that. But I suspect that those that are creating the next version should have a strong knowledge of that but that doesn't mean that they do.
I worked in an office with probably 20 developers and I was the only one that understood bit manipulation. There was one other person that recognized what I was doing but could not explain what I was doing. That didn't make them stupid people it just meant that they had different areas of expertise than what I did.
I worked in one group and being young and inexperienced I was dismissive of another developers programming skills. But once every three months that person, and only that person, would go before a board of people I had never even met to justify what our group did. And every month they allowed us to continue. And the result, which I only learned latter, of failing to do that would be that the group would be disbanded and the members of the group would be required to find another position in the company within two weeks or they would be let go.
Now who was really more important to my actual job, me who I thought I could program or the guy that made sure I actually had a job that I could program at?
|
|
|
|
|
Yeah dude, and I'm on board with that. I'm not on board with crapping on decent software just because you don't understand it.
"Never attribute to malice that which can be explained by stupidity."
- Hanlon's Razor
|
|
|
|
|
You know, there's always a fear to talk about things. Because people put you down instantly.
This whole thread, I've been saying I'm learning things. I'm just conveying what I read & what is there in my mind.
You CANNOT have an idea about a technology/concept, 100% perfect on your mind. Just like how Windows Disk-defragmentation shows you the perfect-blue blocks, all sorted out. It's impossible. Missing blocks-inevitable. You sort this out as and when you get to work on it more, read it more & mainly discuss it more.
Full Reset
|
|
|
|
|
Eddy Vluggen wrote: With most of the world running Windows,
Desktop. Far as I can find server side it is about even.
|
|
|
|
|
Sure, there's a large part running 'nix. Still, aything that is not a hosting-firm will look very closely at the Microsoft-stack, where .NET is thightly integerated with the other products - not just the server products like Exchange and SQL Server, but also Office.
Yes, pointing out the obvious.
Bastard Programmer from Hell 
If you can't read my code, try converting it here[^]
"If you just follow the bacon Eddy, wherever it leads you, then you won't have to think about politics." -- Some Bell.
|
|
|
|
|
Eytukan wrote: What's making Node.js claim this "non-blocking" tag so much?
Perhaps merely marketing or to make those doing it feel special.
But one unintentional (or perhaps intentional - I don't know) is that it emphasizes that you need to insure that your own code isn't written in a way that blocks. Because you can write code that basically blocks everything. So maybe it just acts as a reminder to make sure people figure out the correct way to do it.
Eytukan wrote: IIS is really a dumb, 1-to-1 thread spawning server for every connected client?
Even if so that isn't generally a concern. Everyone likes to think they are writing the next facebook but most are not. And at the point where a company gets big enough that connection concerns are a problem they, probably have so many other architecture problems already that it becomes a non-issue.
|
|
|
|
|
jschell wrote: But one unintentional (or perhaps intentional - I don't know) is that it emphasizes that you need to insure that your own code isn't written in a way that blocks. Because you can write code that basically blocks everything. So maybe it just acts as a reminder to make sure people figure out the correct way to do it.
Exactly, I was truly puzzled how can you "always" ensure non-blocking code.
(edited* - supposed to mean CPU bound tasks. Not I/O bound)
But even with this limitation, How Node.js managed to steal such a huge spot-light? If there's real substance to it, then I would really feel, Microsoft should have done it first , with all the expertise they already have in their hand. Instead of doing a follow-up product.
I've seen Microsoft do this a lot of times. A particular trend gets popular, and they introduce their version of it. You'd call this Innovation? Windows Phone was a typical example.
In recent days, we can really see they are getting closer to being cool
The strength I see at the moment with Node is their NPM, but this is all evolved after the initial hype.
And Microservices architecture really had a good take-off with Node.js. The basic nature of Node solution architecture feels micro. Before Node arrived, "Server" development was an epic. People talk about J2E, .Net N-tier server arch. It did feel like you have a decent learning curve to start doing something on the server. & Deployment , config on the IIS , had it's own learning curve.
Honestly, I feel Node + NoSQL feel bare-bones stuff & anybody can just jump in and start doing things even for a decent requirement, it works. And you'd pay just for the hosting. That's the reason Node became Startups favorite.
jschell wrote: Even if so that isn't generally a concern. Everyone likes to think they are writing the next facebook but most are not. And at the point where a company gets big enough that connection concerns are a problem they, probably have so many other architecture problems already that it becomes a non-issue.
lol
Full Reset
modified 27-Dec-18 13:07pm.
|
|
|
|
|
Eytukan wrote: Honestly, I feel Node + NoSQL feel bare-bones stuff & anybody can just jump in and start doing things even for a decent requirement,
Can you say 'Visual Basic'?
Back in its day Visual Basic contain a vast (for then) array of built in functionality that if one needed to use some other technology one would need to find or more often buy/create to get. So relative programming newcomers could just "jump in" and create their own business application.
People attempted to hoist that into the more complex enterprise spaces and keep it going as more business functionality emerged but it just didn't work.
And Microsoft at point wasn't in the technology business. They were in the application and OS business. They probably only added technology to scare off/kill what they though as competition. (I don't denigrate that process just note that during that time it was not well thought out and definitely was haphazardly marketed.)
|
|
|
|
|
Node gives the same feeling for server what VB did for desktop applications.
You see something getting done with least possible learning curve.
But I cannot say this makes you an expert with technology. It just works.
Full Reset
|
|
|
|
|
I'm creating a report that display construction supplies delivery dates. The supply drops are called "Delivery Packs".
Problems
1. There can be any number of houses being built.
2. There can be any number of packs
Designs
Here is what I considered first. You can see that the Packs could conceivably run off the right side.
So then I considered this. But this doesn't solve the problem as there can be any number of houses, which could run off the right also.
So, I'm not really sure of the right way to lay this out. Any thoughts?
If it's not broken, fix it until it is.
Everything makes sense in someone's mind.
Ya can't fix stupid.
|
|
|
|
|
Does it have to be a grid? I'd be inclined to group by house, and list the relevant packs underneath.
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
ya that's probably doable
Thanks
If it's not broken, fix it until it is.
Everything makes sense in someone's mind.
Ya can't fix stupid.
|
|
|
|
|
But a pack would never be split by several houses, right?
I believe Richard is quite right.
That way you also get a delivery list per house.
|
|
|
|
|
I need to build an application whereby a PC that's already on the LAN needs to communicate with a PC that's newly connected to the LAN.
This is in Windows.
Is there a method in TCP/IP that a PC can use to broadcast, or advertise its machine name and IP address, so that another PC can listen for it and then connect via SSH?
Assume that I cannot use nslookup because in my experiments, it's not reliable.
Thank you.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
I've used "arp -a" to list what a given PC "sees" (static / dynamic IP, MAC address).
The dynamic IP's can then be queried via nslookup for a machine name.
"Broadcasting" implies "listening" (on something) otherwise.
"(I) am amazed to see myself here rather than there ... now rather than then".
― Blaise Pascal
|
|
|
|
|
Thank you very much Gerry.
I'll see if I'm able to make use of that command. 
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
You're welcome!
My ARP "sees" smart TV's etc. without me specifically "looking" for them; so there is some background work (in Windows) going on implicitly. (Same with a router's admin panel.)
(Asynchronously pinging a range of IP addresses on a LAN using a short timeout is another / additional way of looking for new addresses.)
"(I) am amazed to see myself here rather than there ... now rather than then".
― Blaise Pascal
|
|
|
|
|
You can google for "IP Broadcast".
Might want to make sure your needs are appropriate to that though.
For example some times (maybe always now) networks can suppress those.
And it isn't all that efficient nor necessarily even that reliable because it is hard to figure out if it failed because you are using it wrong or because it just didn't get there.
Since it is easier to throw up servers these days that is what people do.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
