TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

The Costs of Programming Language Fragmentation

137 pointsby jp_scover 6 years ago

33 comments

burntsushiover 6 years ago
I also disagree with the premise of this article, and I do not think the argument presented here is compelling at all. The motivations that spur people to do work are complex and diverse, and this article seems to pretend otherwise. For example:<p>&gt; but it&#x27;s common for new languages to trigger reimplementation of, e.g., container data structures, HTTP clients, and random number generators. If the new language did not exist, that effort could have been spent on improving existing libraries or some other useful endeavour.<p>Yes, it <i>could</i> have, but would it? Speaking as someone who has done some of this reimplementation work in a newer language, I can unequivocally say that it would not in my case. I&#x27;m just one data point, but I don&#x27;t think I&#x27;m unusual. There is an aspect of greenfield development that is appealing to me. It is a somewhat unique opportunity to execute a vision with a lot of freedom. There is also the benefit of new forms of expression that new tools give you. I&#x27;m painting with broad strokes here, but I&#x27;m generally a believer in the idea that tools themselves can both limit and empower the expression of ideas.<p>If it weren&#x27;t for the new language, I surely would have done something else with my time. I don&#x27;t know what it would be, but I know it would not be trying to expend the social capital required to make small incremental improvements to existing software using tools that I find too limited. As another commenter mentioned, I might have just watched more Netflix.
评论 #18120925 未加载
评论 #18122701 未加载
评论 #18120644 未加载
评论 #18120202 未加载
评论 #18122907 未加载
评论 #18122748 未加载
foucover 6 years ago
How about the reverse problem?<p>The cost of mindshare. Everyone grouping up on a major programming language or framework. It&#x27;s a bit irritating when people often choose their language or framework because of the size of the community rather than actually evaluating alternative languages or frameworks first.<p>Take React for example. There are some great alternatives to it, especially inspired by react but simpler &amp; cleaner DSLs. But people just keep with React anyways because everyone is already there.
评论 #18119861 未加载
评论 #18122195 未加载
评论 #18119844 未加载
zimbatmover 6 years ago
How much of the new languages are supported by a new fresh wave of engineers that want to learn things?<p>To become a good engineer I had to go through implementing my own container types, play with my own little databases and network libraries, implement a build system, ... I published a few of those experiments in Ruby, the new&#x2F;hip language at the time. Picking a fresh language seems the obvious playground to do this, it&#x27;s where you can leave your mark.<p>My impression was that nodejs developers&#x27; age distribution was quite young as well when it started.<p>My point is that it&#x27;s necessary to introduce new languages once it a while to get good engineers :)
评论 #18121455 未加载
评论 #18122420 未加载
reacwebover 6 years ago
IMHO, the author is completely wrong and the sentence &quot;C is the desert island language&quot; is closer to reality (see <a href="http:&#x2F;&#x2F;www-cs-students.stanford.edu&#x2F;~blynn&#x2F;c&#x2F;intro.html" rel="nofollow">http:&#x2F;&#x2F;www-cs-students.stanford.edu&#x2F;~blynn&#x2F;c&#x2F;intro.html</a>). The fact that such a poor language remains the single sane choice to build the kernel of Linux is a proof of the lack of languages. C++ was already a mess in 98. Python started as simple but has added more and more complex syntaxes. Java is in the hands of Satan, Ada is too verbose, nobody takes the effort to learn it, Haskell and SCALA are niche languages not so easy to learn...<p>IMHO, there is a need for a polyvalent language that would enable development of Linux kernel (low level and performance), that would be easy to learn and that would enable quick development (dynamic typing, syntactic sugar, ...).
评论 #18120884 未加载
评论 #18119651 未加载
评论 #18119940 未加载
评论 #18121657 未加载
评论 #18119833 未加载
评论 #18122525 未加载
评论 #18128846 未加载
flohofwoeover 6 years ago
I really believe that it is better to have many small &#x27;throw-away&#x27; programming languages that can interoperate with each other, than having a few huge languages that are isolated in their own ecosystem.<p>A language shouldn&#x27;t be complex nor should it take more than a few days to learn, and it should be easy to abandon when another language is better suited for a problem.<p>What actually makes many languages useful is their standard libraries and library ecosystem, and the two always get mixed up. When people talk about how great language X is, in most cases they mean how feature-rich or easy to use the standard library of that language is.<p>With few exceptions, those libraries shouldn&#x27;t be tied to a specific language. Let me access your library written in language X from my language Y.<p>Let me easily create projects that are made of different languages, for each part of the project the language that fits this part best.
评论 #18122377 未加载
digitalzombieover 6 years ago
I disagree. They can do whatever they want because it&#x27;s passion driven.<p>If people finds the new language wonderful they&#x27;ll choose to spend their time there and to create their own community around it.<p>It&#x27;s their free time and they have every right to choose how to spend it.<p>Also with RPC, the apache project arrow, etc... there&#x27;ll always be people out there will bridge community.<p>Also is this really a problem? Programming language fragmentation? I&#x27;ve seen front end javascript frameworks fragmentation everywhere and they&#x27;re fine with it. And the solution on the horizon are standardization such as web component (via w3c). These frameworks converge toward web component and now it&#x27;s getting standardize. I&#x27;d argued that because of the fragmentation we actually know what we want to standardize and it also pushed for it.<p>If there weren&#x27;t any fragmentation I&#x27;m not entirely sure if web component would ever be standardize.
评论 #18120141 未加载
评论 #18120621 未加载
评论 #18120362 未加载
TheAceOfHeartsover 6 years ago
I think this could be mitigated in some cases with better tooling. Many programming languages just make it needlessly difficult to interact with anything that&#x27;s not part of their ecosystem.<p>Check out parcel [0], a web application bundler. It has built-in support for lots of different assets [1], with two notable inclusions being ReasonML and Rust! In this blog post [2] they highlight how easy it is to import Rust code from JavaScript.<p>Another neat example is Objective-C bridging on macOS [3]. The code usually doesn&#x27;t end up looking very pretty, and it can be brittle at times, but with JavaScriptObjC you can interact directly with all the native APIs using JavaScript. Here&#x27;s a blog post [4] showing how to write a native app on macOS using JavaScriptObjC.<p>[0] <a href="https:&#x2F;&#x2F;github.com&#x2F;parcel-bundler&#x2F;parcel" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;parcel-bundler&#x2F;parcel</a><p>[1] <a href="https:&#x2F;&#x2F;parceljs.org&#x2F;assets.html" rel="nofollow">https:&#x2F;&#x2F;parceljs.org&#x2F;assets.html</a><p>[2] <a href="https:&#x2F;&#x2F;medium.com&#x2F;@devongovett&#x2F;parcel-v1-5-0-released-source-maps-webassembly-rust-and-more-3a6385e43b95" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;@devongovett&#x2F;parcel-v1-5-0-released-sourc...</a><p>[3] <a href="https:&#x2F;&#x2F;developer.apple.com&#x2F;library&#x2F;archive&#x2F;documentation&#x2F;LanguagesUtilities&#x2F;Conceptual&#x2F;MacAutomationScriptingGuide&#x2F;HowMacScriptingWorks.html#&#x2F;&#x2F;apple_ref&#x2F;doc&#x2F;uid&#x2F;TP40016239-CH73-SW4" rel="nofollow">https:&#x2F;&#x2F;developer.apple.com&#x2F;library&#x2F;archive&#x2F;documentation&#x2F;La...</a><p>[4] <a href="https:&#x2F;&#x2F;tylergaw.com&#x2F;articles&#x2F;building-osx-apps-with-js&#x2F;" rel="nofollow">https:&#x2F;&#x2F;tylergaw.com&#x2F;articles&#x2F;building-osx-apps-with-js&#x2F;</a>
DarkWiiPlayerover 6 years ago
That reflects rather well how I feel about the recent DSL craze, specially in the ruby community. It&#x27;s also how I feel about libraries and frameworks.<p>Programming languages probably suffer the least from this phenomenon due to how difficult it is to create a complete language, with compilers and all.<p>In contrast, languages that just extend others often <i>do</i> benefit from existing communities. Take moonscript¹ for example, it&#x27;s just a new syntax for Lua, so you can use all the existing libraries at no cost. Or take Terra², which can make use of all the C libraries out there <i>and</i> the Lua libraries at the same time.<p>¹ <a href="http:&#x2F;&#x2F;moonscript.org&#x2F;" rel="nofollow">http:&#x2F;&#x2F;moonscript.org&#x2F;</a><p>² <a href="http:&#x2F;&#x2F;terralang.org&#x2F;" rel="nofollow">http:&#x2F;&#x2F;terralang.org&#x2F;</a>
评论 #18119598 未加载
评论 #18120217 未加载
评论 #18119636 未加载
评论 #18120469 未加载
fauigerzigerkover 6 years ago
Isn&#x27;t this a pretty trivial case of exploration vs exploitation? Clearly we need both.<p>The more interesting question for me is why we still don&#x27;t have a widely adopted approach to sharing libraries across languages.<p>The most promising approach I have seen in recent decades was Microsoft&#x27;s COM. But it has declined along with Microsoft&#x27;s clout.<p>Unfortunately, current industry leaders don&#x27;t seem to be interested in this problem at all.
评论 #18120634 未加载
评论 #18122366 未加载
评论 #18121425 未加载
alexmorleyover 6 years ago
I understand the concern but it assumes that open-source contributions are a zero-sum game, which I don&#x27;t think they are. New languages often encourage users to become contributors where they might have remained users in the old language [citation needed]
alexpwover 6 years ago
&quot;If the new language did not exist, that effort could have been spent on improving existing libraries or some other useful endeavour.&quot;<p>This ignores the value and usefulness of learning for the sake of learning, and to allow other people to hone their skills.
评论 #18120408 未加载
amiroucheover 6 years ago
Thankfully most of the comments here on HN argue:<p>a) Your advice on how I spend my free time is not welcome,<p>b) Same ideas that apply to science apply to FLOSS code. You can do and share you research.<p>And completely disagree with the idea that if I publish some code I do a commitment to maintain that code. That&#x27;s written in most (all?) FLOSS license: this software comes with no warranty.
michaelmroseover 6 years ago
&quot;However, I hope people consider carefully the social costs of creating a new programming language especially if it becomes popular, and understand that in some cases creating a popular new language could actually be irresponsible.&quot;<p>Passion isn&#x27;t fungible and there is zero gauruntee that the person who creates an interesting new language wouldn&#x27;t have chosen to binge watch Netflix instead.<p>Furthermore without a crystal ball it&#x27;s difficult to separate ahead of time which efforts will move us forward in some small or large ways.<p>The entire premise of this post is flawed.
评论 #18119431 未加载
评论 #18119567 未加载
cntlzwover 6 years ago
So, if we follow the authors logic we should abandon all science because we don&#x27;t know what they might cause.<p>Innovation is pure by nature. Judgment is done empirical evidence after innovation. You can&#x27;t change the order.
评论 #18122870 未加载
评论 #18123515 未加载
zzzcpanover 6 years ago
Programming languages and libraries are not universally expressive or useful and are not of universally good quality. New independent languages and reimplementations are necessary to better express problems and therefore reduce defects and maintenance costs. They are also necessary for anything of quality to emerge, for all the right people to make all the right choices. Incidentally this is one of the reasons that prevents large projects from achieving good quality and large organizations from producing quality software.
评论 #18120734 未加载
stellaloover 6 years ago
When exactly have diversity and choice become problems?<p>It’s a blessing that so many new languages are popping up. Those that have a reason to exist, will succeed, and the time spent on them will be an investment; those that will fail will do so for a reason. And that’s ok. Some people will have wasted their time. But not the community as a whole.<p>Failures are part of evolution. Thanks to failures we have Rust and Julia and Nim and who knows how many other new interesting languages.
sisciaover 6 years ago
I don&#x27;t agree.<p>With times languages become obsolete, so from one side they try to &quot;upgrade&quot; their syntax, semantic and internal, from the other side they keep backward compatibility.<p>Keep adding features on top of something that was not designed for those features will simply create a huge mess to work with.<p>And arguably the oldest languages are the one where you need to follow &quot;best practise&quot;, where there are thousands of way to achieve the same goal, etc...<p>It is not bad, but we really need newer languages to move forward the field. Most of them will fail, some of them will fail while having used a great amount of resources, but some will eventually succeed and moving the field forward.
评论 #18120691 未加载
PedroBatistaover 6 years ago
Thank god nobody listened to his advice, otherwise we all be doing Basic and Java ( such rebel’s)
ChrisCinelliover 6 years ago
It is not just creating a new language. In a company environment you see a huge cost when people want to start using a new language or&#x2F;and framework.<p>It comes down to values and what push people forward.<p>I am not saying one language should be enough. I am saying that for example that even for a company that already has a library of more than a few components built in React a a medium codebase using it, starting using Vue will add a quite large cost. Even if a framework is 20% better in some metrics, having to rewrite some reusable components (assuming they were written well) in another framework is going to sunk a bunch of time. There is a learning curve that people need to go through. Hardly the first time you use a framework, you end up writing the most awesome code in that framework. There are bugs, there may not be a lot of documentation or open source libraries. From an engineering prospective is going to be fun and you are going to learn something new but from a business prospective it is rarely a good decision.<p>Same thing is probably true when teams in company with a large Java codebase want to switch to Scala.<p>On the other end when you have a 10x better framework or technology you are better paying attention and make sure you plan a transition sooner than later. For example, I think GraphQl is a order of magnitude better than REST when you consider a whole end-to-end system.<p>Engineers, especially the smart ones, get bored quickly. Inventing a new largely adopted language or framework is an irresistible calling from inside. Exploring new territories is also another source of growth. It is all about values and the way we go about satisfying them.
glangdaleover 6 years ago
My &quot;side&quot; take on this is that as long as reimplementation is considered an urgent activity with each new language, we may be privileging languages that are actually <i>better</i> for rebuilding things that are already well understood over languages that are better for &quot;feeling our way through problems&quot;.<p>I&#x27;m not 100% sure of this point, but it feels at least possible that features that make a language good for reimplementing a problem with the structure known in advance might be awkward for exploratory work. I&#x27;ve built a bunch of stuff recently where I had NFI what I was doing as I went along and had to tear up vast tracts of the data structures as I went; it was good to have the fairly loosey-goosey typing of a mix of C&#x2F;C++ and (effectively) asm to do it (while still being able to see performance levels of aggressive optimization, which was also part of the goal).
kazinatorover 6 years ago
Developing TXR is one of the best things I&#x27;ve done. From the perspective of a user, I&#x27;m grateful for its existence.
评论 #18136088 未加载
eeccover 6 years ago
Vendorization. For every language out there there is a consulting firm or corporation selling tooling and support. This allows business strategies aimed at market capture and monopolization.<p>To business, you’re not a general purpose plumber, you’re just a specially trained installer of a particular brand of water boiler.
评论 #18119981 未加载
评论 #18124180 未加载
评论 #18119671 未加载
chubotover 6 years ago
It&#x27;s hard to do, but I would like to see more overhauls of existing languages. I wrote a relevant comment about the approach of Checked C vs. Rust here:<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=17944310" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=17944310</a> (tl;dr Software gets rewritten much less frequently than you think. Rust is going to add to the existing ecosystem in C++ and C, not replace it.)<p>Facebook&#x27;s Hack is another example of this approach -- an overhaul of PHP. Although I guess it led to 2 languages and not one -- it&#x27;s not replacing PHP! Like I said, it&#x27;s hard :)
indymikeover 6 years ago
Most new languages come with an improvement, usually of the non-trivial sort and grow communities and survive on their own merit. Usually, the benefit of a new language exceeds the cost... or it simply gets no traction.
jonduboisover 6 years ago
It&#x27;s annoying how new languages sometimes become popular in spite of not adding any value to the development process. I think part of the problem is that new languages open up a new market which incentivises developers to create open source libraries that will become the foundation of the new ecosystem; it&#x27;s a sure way to become a &#x27;rockstar developer&#x27;; if the language succeeds, you succeed with it; the odds of success are much higher than having to compete with other libraries and frameworks within a well established ecosystem.
lazyjonesover 6 years ago
He‘s right about the cost, but it‘s surely smaller than the cost of bloated protocols, specifications, reference implementations. Nobody except large corporations is able to implement a competitive web browser, for example, regardless of the language. New languages would be much less of an issue with clean API designs and specifications.
hpstrkngover 6 years ago
It&#x27;s like asking authors to not write books unless they know they have a New York Times bestseller on their hands...
honkycatover 6 years ago
The opinions expressed in this blog post are one of those uninteresting, infinitely parroted idioms that people like to assert to seem smart and pragmatic, but are completely idiotic upon further reflection.<p>It&#x27;s similar to &quot;If you rent you are throwing your money away&quot; completely ignoring the realities of how DUMB that statement is. Paying for a house which you must mortgage, pay taxes on, and maintain, but cannot rent out or sell is much more expensive than renting someone else&#x27;s house, and historically the stock market gives greater yields for many investors.<p>First of all, we have not yet written the perfect programming language, and every single language written has pros and cons for various different tasks.<p>Choosing the right language can lead to writing abstractions that make you ABSURDLY more productive than alternative choices. Write RabbitMQ in Javascript instead of Erlang and tell me we only need one language to rule them all.<p>Off the top of my head, I can think of 5 languages that have MASSIVELY improved the software ecosystem over the last 10 years: Golang, Rust, Elixir, Typescript, Elm.<p>&gt; However, I hope people consider carefully the social costs of creating a new programming language especially if it becomes popular, and understand that in some cases creating a popular new language could actually be irresponsible.<p>Unsurprising a self-proclaimed christian programmer takes the moral high ground, and asserts writing a new programming language out of passion, love, or an actual need for a new language to allow for cleaner abstractions for your particular use case can POSSIBLY be unethical.<p>I have a different suggestion: Consider the social costs of parroting idiotic drivel you heard one time and trying to pass it off as actually interesting insight.
michaelfeathersover 6 years ago
In ecology, what the author cites as a problem would be seen as a sign of a healthy robust ecosystem.
preordainedover 6 years ago
I think this is one of the best things about Clojure and similar &quot;hosted&quot; languages...leveraging preexisting huge ecosystems. You really do (as much as one could reasonably expect) get to have your cake and eat it too.
toolsliveover 6 years ago
It&#x27;s a very good question. However, thinking about it, the conclusion should probably be that the cost of continuing to use the same language while there exists another language more suitable to attack your problem is higher.
wwwighamover 6 years ago
Fragmentation yields a distribution of effort which yields theoretically inferior products, compared to the theoretical output of all the individual parts working towards the same goal. (Debatable, but we&#x27;ll assume it.)<p>Centralization yields a lack of competition yields a lack of a drive to improve inferior products yields stagnation and disenfranchisement. (Also debatable, but probably safe to assume here as well.)<p>Startup costs in this area (basic compiler stuff, usually handled by llvm and other varied metacompiler frameworks nowadays) become vanishingly small with time, while the long tail becomes ever larger (libraries, tooling, ecosystem goodness, all are expected now).<p>Given these two observations, when a language becomes popular despite starting in a fragmented ecosystem, it slowly grows, starts to take advantage of network effects, and gains the benefits of growing centralization as it becomes &quot;the defacto choice&quot; in some area (usually at the cost of other players in the space, be they large or small). Eventually, it stagnates, parts of the community become disenfranchised and go and spawn their own languages and variants (taking ideas from their origin, along with their grievances and ideas from other areas with them), and the process begins anew. However because of the growing long tail, each time this cycle takes place, the time between expansion and explosion takes longer and longer.<p>At least, that&#x27;s how I&#x27;ve been led to believe systems like these tend to work.<p>It seems like the call to be _mindful_ of what you&#x27;re doing when you make a language is sensible; if you contribute to the cycle, you&#x27;re contributing to what will eventually be 2 years of long-tail work for what will be the de-facto norm for developer tooling 40 years from now; but that&#x27;s a terrible way to look at it. Wouldn&#x27;t you rather get in closer to the ground floor and be part of the first iterations that set the standards for the iterations which come after? Isn&#x27;t thinking of it any other way just being defeatist? (Since it amounts to concluding that future generations of developers must be better at this than you, and so are worth more early-cycle-iteration time?) So what if it&#x27;s indulging the ego some; if that&#x27;s the cost of improvement, then so be it. The languages we have today wouldn&#x27;t have been made without their predecessors, and the languages of tomorrow won&#x27;t exist without the ones we have today. Is it also equally possible that deciding _not_ to make some new language potentially delaying the progress of the programming language field as a whole? Is making that value judgement in the purview of anything other than hindsight? I can&#x27;t begin to imagine in what ways future languages may improve life as a developer, but if they&#x27;re anything like the stark contrasts we&#x27;ve seen recently in certain PL areas (ie, systems languages), it&#x27;s sure to be exciting.
casper345over 6 years ago
Points to the author, just because one can create doesn&#x27;t mean one should. Should there not be more intentional thought in creating and creations ramifications? &quot;Now I am become Death, the destroyer of worlds.&quot;
评论 #18125043 未加载