Comment by dangus | Hacker News Reader

Comment by dangus | Hacker News Reader

Parent Full thread

I think the level of alarmism is way overblown especially from the author’s perspective coming from Gitpod.

There’s so much focus on how Microsoft has this huge advantage with VSCode and GitHub Codespaces as a cohesive product.

But I think that this entire cloud coding workstation space is rather niche in itself.

I think the average user of VSCode, VSCodium, or whatever other fork has full flexibility to not use Microsoft’s preferred extensions and solutions and customize their environment.

All this stuff about fracturing an ecosystem basically amounts to picking default extensions and I just don’t know if this is one of those “abuse of market power” things yet.

I mean, sure, it can be said that it’s bad how Microsoft is so large that it can single-handedly steer the course of a lot of development trends.

At the same time, I don’t think VSCode being such a good product was something that came out of Microsoft’s dominance. They had a lot of luck with that one, they had the right idea, people, and management at the right time. Microsoft makes plenty of software where their big company big dollar investment hasn’t made them a market leader and they have a laundry list of failures and technologies that are trending downward to show for it.

I would agree with another comment on this thread: it really does seem like a bunch of complaining from some cloud workspace competitors who want to get a free IDE for themselves that works perfectly.

if your competitors can't sell your product, it isn't open source.

But you can sell Vscode if you want? Fork it like VSCodium and sell vsmoneycode. You can't sell all the addons, but this is a business model heaps of people use (open product, closed modules/addons)

>But you can sell Vscode if you want? -> yes you technically can but in practice it's almost impossible to be competitive, since for several major languages the best extension isn't open source. Author called it a "fracture".

Isn’t that true for most software though?

Gimp isn’t better than Photoshop.

LibreOffice isn’t better than Office.

Kdenlive isn’t better than DaVinci Resolve/Adobe Premiere/Final Cut Pro.

Heck, a whole lot of I know consider IntelliJ a whole lot better than VSCode.

Open source isn’t really guaranteed to compete with proprietary software and often doesn’t.

The author calls that fracture, but to me an open source product with extensions will logically have both proprietary paid extensions and open source extensions, just like there are vendors who sell closed-source software for Linux.

If I sell my proprietary application for Linux and I only distribute it for Ubuntu, that doesn’t make Linux “not really open source.”

In my opinion that is normal and expected and isn’t cause for alarm.

No, it isn't true for most software. Firefox and Chrome are better than MSIE or Spyglass Mosaic. TCP/IP is better than CompuServe, AOL, and X.25. Apache and nginx are better than IIS. Perl, Python, and JS are better than PowerBuilder and Visual Basic. GCC is better than Borland C++ and various Unix-vendor C compilers. Even today, GCC is an enormously better C compiler than Visual Studio's, although at least Visual Studio does have a competitive C++ compiler in it. Arduino is better, for most purposes, than various proprietary embedded-vendor IDEs, which themselves mostly use GCC. Linux is better than Unicos, Solaris, and Symbian, though iOS and Microsoft Windows are still competitive.

Open-source software has pretty comprehensively replaced proprietary software throughout the computing stack over the last 30 years mostly by being vastly better. Proprietary software is holding onto footholds in a few places, most of which you listed.

The general strategy for preventing the situation where "an open source product with extensions will logically have both proprietary paid extensions and open source extensions" is copyleft licenses such as the GNU GPL.

Fundamentally, this mindset and flaw in the definition is why open source can't prevail. Because open source is absolutely committed to being unsustainable.

on the contrary; it's absolutely committed to not being dependent on a vendor, because being dependent on a vendor makes software unsustainable

You miss the point. "Open source is not a business model" but also the OSI has made it increasingly clear it doesn't want it to even be possible to make a business model on top of open source software.

That means companies will be, regardless of the desire otherwise, forced into proprietary models, and proprietary development will continue to be the most effective way to put food on the table.

Open source will forever be a niche volunteer thing plus some predatory abuse by monopolies because the OSI is not interested in open source being the default way to make software.

>the OSI has made it increasingly clear it doesn't want it to even be possible to make a business model on top of open source

The OSI's definition of open source was worked out in the 1990s and has changed almost not at all. That stability in the meaning of "open source" (or "OSI-approved", if you want to be precise) is a good thing. If you want to experiment with software licenses that (like open source) don't lock the user in or that (like open source) allow forking, but that are better than open source at allowing the maintainer to earn money, then please do -- just don't call your experiment open source. Coin another term.

It is not some awful tragedy that those who want to experiment with software licenses cannot use the reputation of the "open source" label to market their offerings.

Parenthetically, "nice volunteer thing" is not an accurate description of AOSP, Chromium, Firefox and most Linux desktop software.

> AOSP, Chromium, Firefox

All exploitative projects entirely funded and mostly operated by a monopoly found violating the law or awaiting judgment for such in almost every single first world jurisdiction.

Also neither AOSP nor Chromium are used in the real world in practice, over 99.99% of users use proprietary forks.

In other words, thank you for supporting my statement entirely.

> Linux desktop software

Yeah, definitely "not niche".

>Also neither AOSP nor Chromium are used in the real world in practice, over 99.99% of users use proprietary forks.

I didn't think it needed spelling out, but OK: GrapheneOS would not exist if AOSP didn't have an open-source license. Electron, Microsoft Edge and Brave browser wouldn't exist if Chromium hadn't been released on open-source terms.

Surely it is not your claim that because Google has done bad things, nothing that Google has ever done has ever been positive?

Going back to your previous comment, where you write, "the OSI is not interested in open source being the default way to make software."

Do you actually believe that if the OSI had been smarter or more ethical or had listened to your advice (sent back in time perhaps) then open source might have become the default way to make software?

I consider it obvious that that was never a possibility.

possibly you are a time traveler because in this timeline open source did become the default way to make software

OK, my previous comment was clumsy (and I concede that most of the software on the machine I'm using to write this is open source). What I meant to do with my previous comment was point out that my interlocutor seems to believe that the situation in software is just terrible, which implies that if things had gone different, it could be a lot better. But I'm not able to imagine a realistic evolution in which software ends up in a much better place than the current status quo.

it's been much worse in the past, so we know it can be much worse. probably at the time it was hard to imagine the evolution to where we are now, so maybe it can get much better too. certainly i had no idea 30 years ago where we'd be now

OK, then let's focus on the fact that my interlocutor seems to believe he can see how to make it much better than it is now, which is very different from its being possible, but none of us in this conversation can see how to do it with any confidence.

i think it's reasonable for him to try different things to try to make it better, but i share your lack of confidence in his proposed course of action, particularly since he seems to have such a hazy understanding of what has been tried already ;)

Really making this easy for me. Here we go!

> GrapheneOS would not exist if AOSP didn't have an open-source license

First, Graphene OS is used by... less than 0.01% of Android users, so again, thank you for confirming my point. Furthermore, GrapheneOS is harmful for the ecosystem. Similar to VSCode, proprietary services are essentially gated out from alternatives to the default proprietary Android. And the big problem with Graphene and other AOSP forks, is they waste developer effort while protecting monopolies. Every developer-hour wasted on Graphene is a developer-hour not spent on a truly open mobile OS, and that's sad.

> Electron

We could live without it. Nobody needs a whole browser to ship a single page app. But also...

> Microsoft Edge and Brave browser wouldn't exist

So remember that whole part where I said 99.99% of use is the proprietary forks? Edge is, of course, completely proprietary. (Brave is just a cryptocurrency grift, will focus on Edge here.) Microsoft is a monopoly in several respects still, and Edge adds several anti-features like ads on the start page and shopping "features" which collect your shopping behavior. Furthermore, the switch to Chromium for Edge makes the Internet ecosystem worse. Because one of the largest independent web rendering engines disappeared overnight... further contributing to Google's monopoly stranglehold over the web ecosystem, and it's ability to essentially force standards bodies to comply with it's choices or become irrelevant.

> if the OSI had been smarter or more ethical

I think time travel conjecture is sort of pointless, obviously there are dozens of conditions on when good licenses got approved, what projects adopted them, etc. But if SSPL was approved, as the viral copyleft license it is, less open alternatives would evaporate overnight. Without the risk Amazon would lift your product, wrap it in a proprietary platform, and simultaneously undercut you on cost, I think at least for SaaS, there'd be very little reason for a business to launch without providing their code on fair terms.

All that being said, the discourse around the OSI and open source is toxic enough and compromised enough, I think shifting to new terms is ideal anyways. I saw https://fair.io/ "fair source" recently, which seems quite reasonable, though I'd love to also see an expression of it more focused around labor and sustainability in addition to just business case. Developers deserve to get paid for their work, and I feel the primary goal of open source at this point is to cultivate free labor for the tech monopolies backing the OSI.

the osi was founded by people who built their businesses on open-source software, with whom i talked about it in detail at the time. despite entrenched resistance from, among others, proprietary software vendors and the iranian government, open source became the default way to make software about ten years ago. in large part that's because it's much better at limiting predatory abuse from monopolies than proprietary software is, so large users commonly choose open-source software even when proprietary alternatives look better on paper

i think you're missing not only the point but actually reality as a whole

> But I think that this entire cloud coding workstation space is rather niche in itself.

It's common / popular in certain enterprise circles, for "security" and audit reasons. No admin or most permissions on your computer. You get a specific VM with "controls". Anything required might be installed on request.

A locked down cloud hosted IDE might just be 1000% easier.

I've only seen it used for outsourced staff in big enterprises, but yeah - security and compliance.

Reply on news.ycombinator.com