Fortran Web Framework
fortran.io
fortran.io
JS doesn't have real issues, when I say JS I say in the context in which it was created, AKA the browser. Browsers have an huge API where everything and its contrary are possible.
The problem is Node.js delegating the most basic things to a for profit company, NPM and it was 100% by design... with Node.js, you can't even parse a multipart request without a 3rd party package... the whole "unix philosophy" for packages was purely marketing bullshit and someone got very rich exploiting Node.js bad choices (Isaac...)
Of course you can (if you want to), how do you think all of those 3rd party packages are built in the first place? Using NodeJS APIs of course, that you can also use if you want to.
But why re-invent the wheel when someone already created it? Just make sure the library you include serves one purpose, has a light amount of code, actually does what you want it to and doesn't change their own API willy nilly. Following these guidelines (for any language ecosystem you use) leads to a lot less hassle when it comes to dependencies.
Don't be obtuse, by that logic you don't need to use any third party package to write a professional app backed by database in node.js, just write your own MYSQL/Postgres driver? hey /s
My point is I shouldn't have to download a package or write a multipart request parser to manage files sent to a http server.
Someone made a profit out of that stupid situation with a terrible package manager, NPM, all by design.
Node.js creator himself said that relying on NPM was a terrible mistake, that's why he went on creating DENO.
Honestly, how many languages ship something like a multipart parser with the core API? And to be frank, I don't think I'd like the language I'm using to do this, only ~30% of my projects touch web-related stuff anyways.
> Someone made a profit out of that stupid situation with a terrible package manager, NPM, all by design.
I agree that NPM/NPM Inc is horrible, but for lots of other reasons. Also don't think it was on purpose, just poor and rushed design.
> Node.js creator himself said that relying on NPM was a terrible mistake, that's why he went on creating DENO.
So? Doesn't mean he is right, who knows what have happened with NPM? Maybe NodeJS would never have taken off in the first place. Brendan Eich also apparently doesn't like gay people, does that mean every JS developer needs to think like him?
He is absolutely right, NPM was a grift all along. The whole "unix philosophy" argument to justify a paper thin std lib was a farce. NPM architecture is terrible to begin with. NPM is designed the way it is(was, fetch X times the same package instead of linear dependencies) because NPM corp targeted growth as a startup, not eco-system stability.
But that NPM was a grift, std lib was a farce and everything designed to fuel growth of NPM Inc is gonna need to have some more evidence behind it than you feelings.
And please, I really wish you do have proof of this as I'd like it very much if NPM Inc got put into their place. But I find it unlikely they designed things for this purpose. In the end (at least for me), the quest for truth is more important than what I think is right.
No it doesn't, by any modern language standard (Go, Python, PHP, Java...)
Node.js creator created DENO because he thought relying on NPM was a terrible mistake. If even the creator of Node.js said that, then he knows that a bunch of grifters profited from the bad choices made back then.
Weren't people here outraged at Copilot? Do people really believe that Microsoft isn't running copilot on NPM packages?
I assume you’re talking about the left-pad incident? It happened over 5 years ago and policies have been put in place to stop it happening again. It was a mistake, they learned from it and moved on.
Secondly, this issue isn’t specific to JS or NPM, a few years ago someone deleted their Golang project on GitHub and broke the ecosystem in a similar way. If anything, it’s less likely to happen on NPM today as it acts as a cache
For that reason, your point about Go only further reinforces their point, since Go was (I think) the first language to make importing another project from the web completely trivial, just one line of code.
On the one hand, I think it's bad for a language ecosystem to repeatedly get in the way of doing something that it many cases it does make sense to do, but on the other hand if you think (most) programmers are going to immediately take us to modern programming dependency hell without that, you can start to see it as a kind of silver lining.
- a plethora of details that are implementation specific; - different compilers and platforms that differ in important details;
- as a consequence of above, most code ends up targeting some limited set is compilers and platforms;
- no standard build system to build package management on top of
- the header file system, which allows api declaration and implementation to be decoupled
By the time you’ve narrowed down enough to make these tractable, you might as well use a given operating system’s package management tools as your “C package manager”.
You're also ignoring the fact that Linux isn't the only operating system in the world. Some of us write multi-platform software, and package managers make it so much easier.
The mere fact that I could not build the project on my local machine (windows) and have to login to a Unix (IBM AIX) was baffling. Yep, Unix, not even linux. And because it was a business critical program which worked, business refused to make any drastic changes to it unless they were absolutely necessary, even if it was a bug fix to a frequently occurring issue.
And I maintained it till 2018 when I left that job. I am pretty sure it is still there in same state.
So, I understand the build tool woes, and cross platform issues. however, I understand it to mean that package manager is not feasible in current environment, not that it does not add value. Recently there was a post on zig, which aims for providing a standard build tool for C (& zig), and later, a package manager for C (and zig)
So, I think such efforts are not only beneficial, but will become reality in future
Package: tarball of code.
Manage: download tarball of code and build/install libraries using configure/make. Then tell the compiler where the headers are and the linker where the libraries are. Easy peasy!
(OK, we know it’s not quite _that_ easy in many cases.)
The elephant in the room is that people on HN and elsewhere violently nod about the need for a “proper package manager” when they are actually in disagreement about what a “proper package manager” needs to do. Meanwhile, it turns out that you can do without one, as the C community and all the “non-proper package managers” out there amply demonstrate.
C targets an incredibly broad spectrum of hardware and software platforms, compiler implementations, and OS layouts. That's both blessing and a curse. I think it would be worth at least trying to converge on a "package" system, even if it takes the next 20 years.
I believe that is changing. A quite large number of influential projects have standardized on Meson. That includes systemd, Xorg, PipeWire, GLib, libui, hexchat, pacman, and (almost) all Gnome projects. It seems fair to say that most people (a) like it, and (b) see it as the most likely path forward for build systems.
That sounds like a distro.
C has a terrible dependency management story.
When it became commonplace to depend on third party or system libraries, because of the special status C enjoys in the software infrastructure, OS and distro developers bend over backwards to accomodate C linkage, providing pre-packaged dynamic linking objects.
The status quo is actually rather unfortunate. Several package management tools exist to attempt to address the issue, but there is no de-facto standard and every project makes its own choice.
> The C programming language predates the time when using third party libraries was common.
For that matter, Javascript also predates the time when using third party libraries was common.
C had such a powerful influence on computing that it didn't need to adapt to the world, the world adapted to C.
Of course, there is more going on with C than the use of new "words" introduced through libraries. Even though it is viewed as a portable language, most of that portability is between hardware architectures and operating systems of the same class. Almost everything above that is handled by vendor specific libraries so there was likely less of a drive to standardize on library management. This has negative consequences in today's environment, but it would have taken foresight to predict that outcome.
And you just install packages from your linux distro repository. If you want to distribute your own code, rhen you make a package out of it and that will handle your dependancies for you.
- you are not using Linux
- you need a newer version of the library than your distro provides
- you want to use a package that is not popular enough to be packaged by your distro
2. cmake and autoconf will both let you list dependencies and provide error messages to tell users to install them. It won't auto-install dependencies like a package.json or tell a user where to get the dependencies, but the message that something needs to be installed can be passed to the user.
Via a Dockerfile? ;)
But seriously, this is typically done in the INSTALL.txt installation instruction file for your library, and it’s up to the user to make sure any dependencies are properly installed and configured.
For dependencies that aren’t commonly installed, tools will often include their source and build them from scratch.
It’s a misnomer to say C package management is broken, because it doesn’t have any.
For executables: actual real statically linked binary. You have a single executable containing all the dependencies and all necessary runtime support.
For libraries: no standard mechanism. On Linux, you can generally make a "libfoo" package that specifies the dependencies. On Windows, this is a bit more of a problem, but people can and do just hand out a DLL.
Or you make people build the dependencies from source as well.
(As an old school C developer, I see containers as occupying the same conceptual space as statically linked binaries for distribution. It's just that very few other languages will let you make a true standalone executable.)
https://cmake.org/cmake/help/latest/command/find_package.htm...
With complex software, just grabbing the latest version of every dependency will lead to broken builds (or ones that only work on x64 but are broken on other platforms)
Last I looked at GUIX (which I guess is similar) you are tied to a release "version" of the package-manager with whatever package version it comes with. But maybe I misunderstood
NixOS (the distribution) is not actually rolling: it has stable channels that are released twice a year, and an "unstable" channel that normally updates every few days (depends when all tests passed). You could call the latter "rolling", but it's a bit different compared to actually rolling distributions like Arch Linux.
> Can you designate specific package versions for all your dependencies? Like having a very new openssl and an older zlib version?
Normally only the last version is packaged, but you can change the version of the package and install multiple versions of the same library without conflicts. Some big libraries have multiple branches for backward compatibility (eg. boost175, boost174, openssl_1_1), for the others you have to override the package (it's pretty easy, but you'll likely lose the binary cache).
> With complex software, just grabbing the latest version of every dependency will lead to broken builds (or ones that only work on x64 but are broken on other platforms)
If you really need that kind of stability, a common solution is to pin the version of the whole Nixpkgs. Something a little less drastic is simply using the stable channel, which only receives security updates.
> Last I looked at GUIX (which I guess is similar) you are tied to a release "version" of the package-manager with whatever package version it comes with. But maybe I misunderstood
I'm not familiar with Guix, but I think you can do what I described above with it too.
I mixed things up a bit. Guix doesn't have a stable channel but Nix does. Twice a year sounds quite fast - but maybe that's just what I'm used to from Ubuntu LTS. I can't imagine being a package maintainer and having to deal with build issues that often (I guess it only targets x64 so it's not too bad)
> If you really need that kind of stability
I guess I don't really 'need' it .. but I just rather my system not be constantly updating and changing under me.
Looks like you have run stable and add unstable packages as needed:
https://discourse.nixos.org/t/installing-only-a-single-packa...
(like running LTS and adding PPAs)
Guix has Inferiors [0] for this nowadays.
There's also a lighter-weight but less robust approach: define a new package inheriting the package in question, changing only the version number and source URI.
And then there is specifying linking files and include files, and headers, it makes it so challenging to even run other people’s code
There is a bit of learning curve but everything is there for a reason.
- Toolchain files define your target system and build toolchain.
- Hunter specifies your dependencies withing CMake (using git and hashes) and does proper incapsulation/namespacing (which isn't present in "normal" CMake) as well as setting up the proper "inheritance" of headers and stuff.
- CMake does the final linking of build artifacts and headers and does the final build command generation.
A proper setup can build both dynamic and static builds for any system with virtually no platform specific switches
Plus many of the implementations are closed source.
C also tries to make as few assumptions about the platform as possible. It was only recently that support for ones-complement arithmetic was dropped http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2218.htm / despite never being used for decades https://stackoverflow.com/questions/12276957/are-there-any-n...
C is also somewhat unique in supporting cross-targeting, where you build a program on platform A that will run on platform B (such as a microcontroller), and the target platform may be very different (such as no filesystem, no OS, different instruction set). Most other modern languages use intermediate bytecodes and try to be "WORA" (write once run anywhere).
For instance, what is this if not package management?
apt install libpng-dev
This is the problem that npm, pip, and whatnot solve. They work the same, (almost) everywhere.
C/C++ don’t have a proper answer for it.
https://news.ycombinator.com/item?id=11938405 (227 points by mapmeld on June 20, 2016 | 90 comments)
https://news.ycombinator.com/item?id=13226174 (165 points by da02 on Dec 21, 2016 | 113 comments)
https://news.ycombinator.com/item?id=22120285 (422 points by lelf on Jan 22, 2020 | 247 comments)
But I would’t use a Fortran web framework. It’s not well-suited to that.
Fortran is one of only two languages with a concise array syntax that’s used at the most demanding levels of high performance computing. The other is Julia.
In situations like that, the effort to make it so that you can build the Web-facing parts in a more Web-friendly language is often far greater than the effort to just build the bits you need in the main language.
(This has been a repeated pattern in Fortran since then, too, unfortunately. J3 invents something, often with flaws that go unnoticed until somebody tries to implement it; compilers don't invest in as-yet unused features, especially flawed ones; codes don't use them because they're not portable, or not performance-portable; repeat. DO CONCURRENT is the latest example -- J3 defined it as a serial construct and included semantics that can preclude parallelization...)
The second one was extra fun because the library was actually written in "FORCE", an extensive macro-extension of Fortran, for which we no longer had a compiler. All I had was docs, the original published paper, the original source, and a handful of input-output examples on substantial datasets
(for context, the datasets were essentially a 1024 x 1024 x 1024 array of floats that needed to be convolved with two different Gaussian kernels and compared pixel by pixel in an interesting way).
(Yes I know it’s simulink for many, but I’ve met some pure computational folk who swear by matlab as their prototyping tool and python is a bug ridden rat nest of footguns to them.)
C’est la vie
By no means am I suggesting that it wouldn't have an advantage; I'm just a young whippersnapper who has never had the chance to write any Fortran.
Thinking about it, I'd say the comparison to Rust for mastery of its particular domain is actually quite apt.
I cut my teeth on Matlab, and flipped a coin for Python or Fortran as my language to focus on after Matlab tried to charge our HPC center per core.
Python won the toss, but I've always had a love for Fortran.
There's a lot to learn in higher level physics, so having the simplest language, without tons of features to fiddle with, while also being very, very performant is preferable.
I have a feeling that much of this work end end up in Julia however.
They use % as the accessor instead of . (Yes I am joking around) ;)
FWIW...Fortran has had explicit OOP features since at least the 2003 standard (e.g. "type extends"). You aren't "required" to learn those features to use Fortran, but that's true of any number of ostensibly OOP languages.
Lastly...endless people (virtually always non-physicists who wrote a couple of lines of Fortran-77 in college) are continuously popping into the conversation with "well of course dump musty old Fortran for the NewHotness language because ew Fortran". Hasn't happened yet, and the Fortran folks have evolved from -77 to -90, -95, -2003, -2008 and lately -2018, so why would it?
I’m wondering why Fortran is dominant in physics yet APL in finance.
I remember writing some data analysis code in Fortran for my freshman physics lab and the TA was surprised to see my choice of language.
Pretty much everything you would ever want to do related to partial differential equations or computational fluid dynamics has already been written in FORTRAN, and is damn good code.
It helps the the language itself is more than capable of great perforamance and that NASA has poured massive amounts of cash into making sure it stays performant on newer hardware.
Rust is also pervasively noalias, but it's taken years for it to actually enable "noalias" on LLVM. Rust code uses that feature more than it's ever been used before, which keeps exposing codegen bugs that no one noticed before. Once those get flushed out, I still expect it will take a while for the resulting optimizations to reach the quality & maturity that Fortran has had.
From the noalias = no panacea
To the Fortran repl
Thank you. This was inspiring.
Maybe it’s time I learn me some llvm too. Cheers.
C has some similar legacy but came a little later to the game and wasn't as intuitive to people driving the money in computing as they were with FORTRAN. It's also heavily optimizeable from a performance perspective but often missed the initial buy in and momentum, though developed much of its own.
So FORTRAN has a lot of investment and strategic advantages to be used for things like weather modeling, at least for specific underlying libraries. Modern work tends not to start from greenfield in FORTRAN, it often starts in C or really a high language like Python to provide theoretical proof of concept and just write/use wrappers to highly optimized codebases like those for FORTRAN for the numeric fundamentals in the model. Lots of scientific glue code these days with not a lot of stuff making it into optimization levels like you see in BLAS, LAPACK, ScaLAPACK, etc.
Python and matlab being to slow at “low barrier to entry level” for sure anyway.
I’m not going to comment on julia because I’m still drinking coffee. In theory… blah blah maybe it should fill this role.
Sit down, write stupid simple Fortran code to solve a large linear algebra problem (a discretized partial differential equation system) and things like matrix multiplication and numpy style mat(:) access are built in. (Numpy borrowed the syntax from fortran, actually)
For a scientist, this is way easier than wading into c++ and getting it right; or sticking with c, and avoiding the footguns.
Fortran gives you maximum serial speed with minimum effort… it is easy to write efficient serial code, while knowing almost nothing of programming. Memory management sure, but even there, there out of the box are no pointers to fumble and your allocations are going to be contiguous. Optimal Array Memory access comes down to knowing Fortran is column based.
Ah but then you need to go parallel, where the documentation more in c these days.
But you’re already comfy with Fortran now, so, scientist that you are, you dig and experiment until you get it running in openmp, mpi, and cuda, etc.
Another point: This thinking cheapens learning. If you only learn, "what isn't dead", you stifle your curiosity and don't learn the whole picture of a thing.
Personally, back in the days, I never really got into the NAG libraries. I found it easy enough to roll my own. Maybe some of the stuff can save you time, but I ended up preferring hand-coding my algorithms.
The greatest thing to happen to Fortran was Fortran 90. None of that column malarky. Modules was kinda OK, but if I wanted to use well-structured code I probably wouldn't be starting with Fortran anyway. I don't think ANYONE I knew at the time used to new features other than format-free.
I tried poking around with allocatable arrays at one point, but didn't like it. I guess it was Fortran's way of trying to be like C.
One thing that is rarely discussed is that a programming language isn't just specification, but it is a culture and philosophy shaped by the programmers themselves. One guy made a reference to Cobol, and how object-orientation was an unused feature. He said that what the designers failed to consider is that Cobol programmers just don't "do" object-orientation.
Fortran - even Fortran 77 (IIRC) has some nice little features, like being able to specify parameters in a separate file, and read them in one line. I doubt most of the other guys in the faculty were aware of that feature at the time.
And my standard anecdote ... when I went to a new job, they actually had a little bit of programming in Fortran. We had Visual Fortran. I had to set up the environment to fiddle with something for a client. Set-up was straight-forward. I opened the project file, and the things opened and compiled without a single hitch. I was shocked - shocked I tell you - at how simple the set-up was. Normally one would expect endless futzing around to get the programming environment and libraries in place.
But that's not the real anecdote. One day I was passing by a meeting room. I overheard an outside consultant in discussion with some of our guys about a replacement for Fortran. He was going on about how flexible the system would be, and hey, it you needed to set up extra parameters you could always add them to an XML configuration file he was proposing.
And I thought to myself ... my God, that's complicated. In Fortran, you can read an array into a file in one statement. Bam! His method would have involved external libraries and some serious effort to get going.
I always joke that people should be forced to write Fortran for at least a year. For those that want to be JavaScript developers, two years. That should make people think much more simply about what it is you're trying to do.
A year ago I added programming microcontrollers to my list of things programmers should be forced to do. A resource-constrained environment should focus their minds a little more.
This is what Fortran's namelists are for.
These days, scientific computing is a relatively minor niche, because people have found so many other things to do with computers that science is just one of many use cases.
Fortran was aimed at a specific audience, scientists and engineers, and that audience seems - AFAICT - happy to keep using it. One may consider it an impressive success story.
Personally, I have never used Fortran for anything beyond the hello-world-level, so I have no strong opinion on the language as such; I do know it has evolved a lot, probably more so than C.
The only reason Fortran is still around is institutional inertia.
This looks, let's say, less obviously whimsical. Is there a serious use case for this, or is this another case of "becase we can"? Honest question.
Of course, there's a C implementation and bindings are only ever a few lines away...
If hipsters come in and take that from me I'll have to learn something even worse
WATFOR it..
On a serious note, at least right now it is quite a niche language that appears to have more demand than supply, so might still be a good idea. Even better in fact if its at least attempting to be modern
LAPACK upstream wont move to another language, unless there are significant advantages: https://en.wikipedia.org/wiki/LAPACK And this is not in sight (yet) for operations on block-data like matrices.
Languages don’t get no deader than that.