Live data from Hacker News

Announcing .NET Core 1.0

blogs.msdn.microsoft.com

301–310 of 327 posts

Re: Announcing .NET Core 1.0

#301
post #287

Earlier quoted context omitted.

> The results of the 2016 Stack Overflow Developer Survey are in! F# came out as the single most highly paid tech worldwide and is amongst the third top paying techs in the US https://fsharp.tv/gazettes/f-the-most-highly-paid-tech-world...

Yes and it's clear from what the F# team has said that the hope they have for the future of the language is that the community will invest. That's fine, but let's not think that this will produce tooling anywhere nearly as refined as C#'s stuff. Take F# interactive v the C# one. F# has like a decade lead. Yet the C# interactive editor is smooth, polished, even has VS project integration, something the F# team had tho…

The finance industry where F# shines is willing to invest, I suppose. OCaml e.g. is backed by Jane Street and F# is even simpler in that regard because the hardest part is efficient runtime which is for F# is .NET - already done well by MS. But what some corp willing to invest in tooling for F# could do while MSVS is closed source and is not most transparent IDE in the world to say the least wrt plugins.

> too proud to admit their flagship language from their high-profile hire

That's acute while we know that high-profile hire's past victories (Turbo Pascal, Object Pascal, Delphi) were never about the language, but about incredibly polished IDE, compiler, libraries and runtime.

> yet it's still a novelty for, as MS has said "scientific and engineering" applications

MS marketing wisdom is overrated, to say the very least. Look at Tablet PC. Windows XP Tablet PC Edition released in 2001. Ink APIs are all way through Windows SDK since back then. And they never realized that stylus thing is something more than just a 'uhm, you can draw a kitten, maybe?'. At least they never articulate anything more than that in any promotion campaign. And now Apple released iPad Pro and will eat the TabletPC-Covertable-Surface market.

Re: Announcing .NET Core 1.0

#302
post #4

After so many years microsoft finally presenting their tools in a more moden way. Never been a huge fan of .NET but i can't deny is a great tool hopefully people try it out on more platforms.

Have you tried writing a complex wep app/api with C#? It's an absolute joy to use. Python still has it beat for math and ML libraries, but for getting up a quick web app with robust (and easy) data queries (LINQ) it's pretty unbeatable.

Re: Announcing .NET Core 1.0

#303
post #302
post #4

After so many years microsoft finally presenting their tools in a more moden way. Never been a huge fan of .NET but i can't deny is a great tool hopefully people try it out on more platforms.

Have you tried writing a complex wep app/api with C#? It's an absolute joy to use. Python still has it beat for math and ML libraries, but for getting up a quick web app with robust (and easy) data queries (LINQ) it's pretty unbeatable.

Yes, LINQ is really good. With these news i am willing to go back and try it in some more project.

Re: Announcing .NET Core 1.0

#304
post #291

Earlier quoted context omitted.

How, exactly, are you thinking that Microsoft is going to fix nVidia's buggy drivers? They can collect all the data they want, but at the end of the day, it's nVidia's driver.

I worked there. Ways we solved these sorts of problems include: hardening the other side of the API/HAL when appropriate/possible, simplifying the driver model so that mere mortals could write drivers, writing our own drivers and overwriting known buggy ones for companies that couldn't get their shit together (usually network vendors), adding workarounds to the OS not to use certain features of certain cards, flying…

Nice work, thanks for the insight!

Re: Announcing .NET Core 1.0

#305
post #162

If I'm an experienced Python developer who develops a lot of websites and APIs using Django or Flask and SQLAlchemy, is there any reason to try this stuff? C# is a great language, but what about the libraries? What replaces Flask? What replaces SQLAlchemy? How do I deploy? Looking for some practical reasons to invest time on this if Windows development is not in my roadmap.

Well, besides libraries, performance. .NET supports proper multithreading (no GIL) and it is also much faster than CPython. If IronPython gets ported to .NET Core (maybe it already has, I don't know), then you'd get those benefits for free.

The performance of IronPython was horrible and Python being an ecosystem of libraries built in C and wrapped in a Python API, it makes alternatives to be unusable, since most Python libraries that matter are incompatible.

Re: Announcing .NET Core 1.0

#306

Earlier quoted context omitted.

False dichotomy, you can use both together and get the benefits of both an ORM and raw sql, each when they're the better choice.

Certainly. Yet there's been a tendency later to declare SQL wrong and deprecated and to incite everyone to forget about it and go with an ORM as the only way. I do not accept that. The stronger the ORM zealots are pushing the more repulsive the idea of an ORM becomes to me. But of course it's just another instrument that has its place among the others.

Only among the inexperienced. Skilled people use both as needed, even at the same time. You can keep the complex SQL in a view and map the view with the ORM. ORM's are far too valuable to ever hand write all that SQL and SQL is far too flexible to ever ORM 100% of every use case, the ORM should make up most of the program with hand written SQL sprinkled in where useful when the ORM is clunky.

Re: Announcing .NET Core 1.0

#307
post #171

Earlier quoted context omitted.

fwiw against Java -- http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan... against Node.js -- http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan... against Python -- http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan...

I find the binary-tree benchmark specially disturbing, it looks like the .NET Core GC/escape-analysis has some catching up to do.

The multi core difference?

Check the cpu secs and cpu load; check the source code.

Re: Announcing .NET Core 1.0

#308
post #171
post #28

I'd LOVE to see some Linux web/micro benchmarks... particularly against Java, Go, Node, Python.

fwiw against Java -- http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan... against Node.js -- http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan... against Python -- http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan...

Missed one:

against Go --

http://benchmarksgame.alioth.debian.org/u64q/compare.php?lan...

Re: Announcing .NET Core 1.0

#309

Earlier quoted context omitted.

Does this count? https://www.visualstudio.com/en-us/products/code-vs.aspx

I've played with VS Code a bit, it's a pretty cool tool for small scale scripting, but it's not really a replacement for a full fledged IDE. Hopefully Jetbrains will release their cross platform C# IDE sometime soon. I would prefer Visual Studio, but I don't think that's going to happen (at least not anytime soon).

Not being fully fledged IDE can be good and bad. I'm using VSCode for most things I use VS for. It's super fast to launch, great language extensions, debugger, integrated terminal and tasks. It's like sublime and visual studio had a baby. I have licenses for jet brains but I seldom ever use it.

I may also be a bit biased since I contribute to VScode extensions. But I see that as a positive. I use a free editor that I can hack, look at its source and collaborate in the open.

Re: Announcing .NET Core 1.0

#310

F# doesnt work ootb with sdk preview2. Work ok with preview1 of SDK (win/ubuntu/osx/docker), but preview2 has a bug, the fix is in progress ( ref https://github.com/dotnet/netcorecli-fsc/issues/12 ) and will be published a nuget package with fix soon.

I have just run into this problem when migrating my project. I see the build is nearly done; when it is, what's the best way of patching?
Post reply on HN