I prefer privately hosted web and mail servers. Before "the cloud", the economy worked just fine and companies had enough money for in-house IT.
By using the nextcloud example, the author of the article is asking the wrong question.
Yet they dont.
The problem is not big tech. Not open source. It's that the European tech economy crippled itself and cries wolf about it all day.
It's the same kind of political denial as with being "concerned" about climate change, but still trading with China (even worse when this allows for fake self-congratulations about decreasing greenhouse gas emissions, when it mostly comes from most of the industry having been exported).
The USA is enjoying the wealth they gained in both world wars and they also kept their defacto colonies in the South America and Pacific.
Europeans destroyed their wealth in the world wars and they lost their colonies. Of course the end of the colonialism has ended some of the human suffering but it has a cold-hearted economic impact.
The American venture capitalists are all coming from the industries that got stronger at and after the world wars. They invested silicon and then the tech industries that built the wealth exponentially. The Europeans had to rebuild their countries until 70s and the investments they made are smaller. Similarly the US spent its government money to nuclear and space programs that further strengthened the economy. EU spent its surplus to improve post-Soviet countries which may or may not pay dividends in the future.
It may require significant reallocation of resources from certain places to tech. It may require diverting the resources spent on old pensioners who are the biggest voting block. It is not a simple lack of political will. It requires reshaping a century of decisions.
A necessary, not sufficient, piece of the alternative
We need to work on our orginistaional structures
What our educational alternatives show, and they have been implemented in some places and in Greenland I believe. Is very much in line with what the article recommends at the end, as far as small incremental useful changes with clear and cut goals. What would you achieve with Nextcloud? Replacing everything you have in Azure AWS in one big step? Obviously that is going to go horribly. That’s not even how we migrated into Azure from on prem. What you can do, is to start by slowly moving your applications and services into moveable parts, by container rising them. Writing your run-books in Python rather than Powershell and so on.
Then there is the change management, which the article touches on, and which is always forgotten by decision makers. Partly because decision makers don’t know what IT is, well… I guess that is it really. Where in the past (and I’ve written about this a lot) SysAdmins and supporters were unlikely to want to leave their Microsoft training, I think we’re at a point in IT history where that is less of a case because so much is now done on Linux even if you’re deep into the Microsoft ecosystem. Similarity the Office365 platform is not in as much ownership of your employee base because many people under 30 will not have “grown up” with it. Where it would have been inconceivable to not use Word, Excel, PowerPoint or Outlook 5-10 years ago we’ve entered a world where we actively have to train employees in Office products because they are used to iOS, Android and MacOS and not “PCs”.
Again, you should start by doing things in small steps. Our Libraries have switched to Ubuntu on every public PC, and it has been a non-issue because many library users are equally unfamiliar with Ubuntu and Windows, and since most things happen in a browser anyway, the underlying OS isn’t an issue.
That is how you do it. Slowly with small steps, and yes, some of those steps don’t need to be open source. If you want to replace Azure or AWS then it’s much better to head to Hetzner (or similar) rather than to try and do it with NextCloud or similar. Because then your SysAdmins will not really need much retraining as that is not very different from what they already do in many cases where moving into the cloud has really just been moving a bunch of VMs.
More news at 11.
>> (By the way, all new software without accompanying support & guidance is doomed to fail. And if that software comes from a dominant player, you’ll just have to deal with that by the way.)
There's a temptation to conflate the software license with the software business. This is natural, but places software as the primary value in the chain.
From a business perspective the software though is a cheap part of the chain. And the least interesting part.
I don't pick say accounting software based on price. Or access to the source code. I base it on effectiveness. And a big part of that effectiveness is that staff can run it. And when it all goes wrong there's someone to call. I'm buying a -relationship-, not software.
Thats why RedHat is a business. They're not selling Linux, they're selling the reliability, longevity, services, support etc.
In truth the license doesn't matter. My accounting software might be open or closed. My supplier doesn't sell me based on the license. They sell me by convincing me that everything just works, and when it doesn't they'll be there to fix it.
>In truth the license doesn't matter.
It's funny to bring that up in the context of Red Hat who have started to circumvent the GPL by terminating their relationship with anyone who tries to actually make use of the rights granted by it. "The license doesn't matter" because they've found a loophole in it, but it clearly does matter in that they had to do so in the first place and weren't able to adhere to its spirit due to business concerns.
[1]: https://sfconservancy.org/blog/2023/jun/23/rhel-gpl-analysis...
[2]: https://opencoreventures.com/blog/2023-08-redhat-gets-around...
This is only because true most of the time businesses use a lot of publicly funded work without paying for it. If software development were entirely private, I'm sure businesses would find excuses that actually no it has to cost 100x what it would cost otherwise.
Everything you say about maintainability and stability is true. But writing software that can be operated as a service in the first place is substantially harder. It's just not as easy for a company to capture.
and they'd tell you to pay up 10x, or lose this stability in the future;
If it was an open source software, you will have the option to go to a competing vendor.
You could say that Canonical and IBM RedHat compete on offering Linux support, but the reality is that it's not that much harder to switch from RHEL to Ubuntu than switching to any other OS, so I don't think this counts.
You miss the point. Enterprises don't go looking for another vendor. Vendors come to them with a sales offering.
If I'm running SQL Server the I pretty much know where I stand with Microsoft, and there are endless MS approved support people.
With PostgreSQL some vendor has to come to me and convince me to switch. PostgreSQL is really well supported, and it's at least an option. 99% of Open Source though has 1 or 0 support entities, and 0 sales people.
Sure, with PostgreSQL I can do my own research. I might even have skills to do it myself. But now I have to explain my choices all the way up the ladder.
Am I going to use an OSS accounting system with no sales people? With no support people? Or am I gonna pay $99 a year or whatever for QuickBooks?
I totally agree with this. And not just businesses, individuals too.
On the flip side lot of open source devs are going to get 100x more productive in the Exploit part than the avg coder monkey at large corp.
Nothing is obvious and predictable about where that story goes in an ever growing ever changing system.
Large corps will keep funding whoever gets the job done. While AI might replace lot of Large Corps activity which is basically on the Exploit side of the Tradeoff.
It’s hard but I still think that’s the way to support OSS
The rise of the Internet and the dot-com boom happened largely without OSS, on proprietary UNIXes, proprietary web server engines, and proprietary database engines.
FAANG and other high tech businesses can easily afford very expensive servers and datacenters to house them thanks to the very very fat profit margins. They can also easily afford the cost of an OS license and other software tools.
> In truth the license doesn't matter.
Come on. What matters is the way the business extracts value from you, and the license is part of that. Especially when the software you produce is so great that nobody needs to be called, because it just works.
Still, the licence doesn't matter - while probably being a bit of an overstatement - is somewhat true. If my enterprise relies on an Adobe service, it's primarily about my relationship with them, not the product license.
... But of course, product price and therefore revenue will decline if competitors can sell my product too or customers can download and use it for free.
So you would pick a software costing 1 million over a software that is 90% as effective but costs 1 thousand?
If it fits your budget, and a commercial product has a good sales team (vs a cheaper opensource one with zero marketing), the commercial product is gonna get chosen even if it costs infinitely more. That's basically IBM and Oracle's play book.
The license matters indirectly: if it's open source, you know that as a fall-back other suppliers might be able to step up and take over, if your original guys fail or get too insufferable.
RedHat providing OSS licensed software is _less_ risk than RedHat providing proprietary closed source operating system.
It only doesn't matter if you don't care at all about software supply chain risks.
This is not a sane position in 2024 to hold.
IBM to save it's business had to merge with Red hat almost 50% 50% in 2018.
Microsoft it's security and cloud offering had to, open source it's .net framework, aquire GitHub, ditch Visual Studio fot Visual Studio Code,
ARM is eating the world, it over hauled the x86_x64 architecture, and became the Defacto architecture.
We can go on and on and on and on,that the Open Source business model, became necessary to survive in tech, not just to exist.
If you don't open it, they will eat you up.
Meanwhile, we continue to pour money into Oracle licenses, not just for basic access but for additional features—like enabling data reading and analysis on the Oracle-embedded database in our main app. And, if we need to allocate more CPU cores on our VMs, we face yet another round of licensing fees.
Sometimes you don’t need much support. Yet pay tons of money.
Why was PostgreSQL not an option according to management? I would not take their dismissal at face value. I'd want to know why not. But that might be Dutch culture.
A good example is the GIS industry where ESRI (ArcGIS) dominates. In Europe the open source qGIS is generally an acceptable alternative despite less 'support'. In America its hard to find anyone using qGIS and ESRI is basically a monopoly.
> An Open Source experiment meanwhile is typically operated by an enthusiastic hobbyist with borrowed equipment. Rolled out without training and without professional support, by someone who likely did this for the first time, it’s no wonder things often don’t work out well.
> After the experiment, the faction was disappointed and concluded that Nextcloud was no good. And that was also their lived experience. “Let’s not do that again!”
This is a rhetorical trick known as implication or insinuation. By presenting information indirectly, the author prompts readers to make a connection themselves without explicitly stating it.
The author implies that the European Parliament's failed experiment with Nextcloud was due to a lack of professional resources and expertise, suggesting it was handled similarly to typical open-source projects led by hobbyists without proper support. However, he doesn’t provide any factual evidence that the Parliament’s Nextcloud experiment actually lacked professional resources, training, or adequate equipment. Instead, he hints at this by describing common issues with open-source setups, leaving readers to assume the experiment suffered from similar shortcomings.
I would have appreciated some facts, or even sources for his claims, but there are none. And I couldn't find any information about the Nextcloud deployment having failed.
I hate to break it to you but it takes time to implement closed source solutions as well. They also always have terrible documentation, because they make money on support.
Purely open source stuff lives and dies on how easy it is to start up.
Closed source paid stuff doesn't need to be easy. Often a decision has been made before implementation, and there are people to help you through it.
It's also easier to get approval for open source most of the time because there isnt a new bill, just my time.
I usually reach for open source first.
[1]https://en.wikipedia.org/wiki/LiMux
[2]https://arstechnica.com/information-technology/2024/04/germa...
Hard to be an alternative when you serve the same master.
Then the whole issue with non-copyleft licenses, that are nothing other than the old Whateverware or Public Domain licenses from the 16 bit home computer days.
We already had access to source code back then.
And for a large crowd this is already good enough, they aren't into it for religious definitions.
I remind you all Emacs powered some German airline's ATC in the early 90's, and it used to be used under Amazon for tons of stuff thanks to its easy widget UI to achieve tasks with very little Elisp.
You can use Google docs for free so it takes some dedication to self host that and pay for the server.
Now if big tech charged for everything things would be more like the old days where you might use small tech, such as a local hosting provider that does open source installs.
every now and then open source is suggested as superior, because being free. Zero comment on code quality, who wrote it, why it came to be in the first place.
Even the argument that a host running open source makes delivery more trustworthy is super biased - major cognitive dissonance is that services based on open tech are very often not open, neither auditable.
There’s a lot of open source being controlled by same large corporations and the part that is not, does not constitute a service on its own.
Then we must admit it takes a lot of care taking care of services nobody else cares about (by means of support).
While open source is important for academia, I think open results are more important for government. Like I don’t care what somebody used to cater to this geospatial data, or that image. I care about the data that went in and went out. Open data is much more important in the era of open weights and closed sources training sets.
The general public is often misled to equate open source to free beer. Well that is also not entirely correct given plethora of not so free licenses. Asp not correct as costs are greater when you put the personnel running that service in the equation. I can see how this argument does not fly well with socialist ideologies, but that’s problem of ideology, not of costs or technology.
Even if we consider only those open projects which are also free - these come with less guaranties than a pair of second hand shoes bought from random store.
Don’t get me wrong - open source is great and we use it daily, but comparing means of distribution with quality of service is really like comparing ябълки и круши (apples and pears in Bulgarian). So it’s indeed time to stop blindly waving the open source flag, but actually try to understand the benefits and challenges it comes with.
almost the entire world and industry is literally running on open source.
Open source software is the building blocks used by large rent (service fee) seeking corporations. They will extract large profits from any of these contracts and that is a demonstrable fact, they are also nearly all from the USA and so those profits will flow in one particular direction. It is also a historical fact that governments have run successful large scale infrastructure. Make your choice.
This logic doesn’t really hold when it comes to large governments. Their needs are large enough that they can justify employing specialists. At that point, the profit margin the service business is capturing is just inefficiency. Internal services should be more common in large governments.
In practice, Microsoft isn't going anywhere. You're just paying for an external inhouse.
imho the question should be if the country continues to function if the project goes bankrupt. If it is so essential that it needs to be saved by the government (even in theory) then it lives outside the domain of capitalism.
Most of my Compose files contain 2 services (1 for app + 1 for database), but some contain 3 and some contain 1. It's incredibly easy to install new software and incredibly easy to shut it down temporarily if you don't need to use it all day.
I'd even argue that some companies would benefit more from using Swarm than Kubernetes. There is a lot of things to take into account when using Kubernetes (let alone setting it up for GitOps), but Docker Swarm can be managed by 1 person.
[1]: A joke, obviously, but it really isn't popular at all
I found setting up gitops via flux quite easy, apart from order of operations, like installing controllers and custom resource definitions before resources that need those CRDs etc.
What were you thinking of things to take into account for k8s over swarm?
The main difference for me is k8s needs a hell of a lot more boilerplate yaml for doing basically anything.
I would be very curious to know if the data are stored on their own data center or Microsoft's.
- [0] https://www.edps.europa.eu/press-publications/press-news/pre...
Try to mimicking them is a waste of time and can't work, pushing the society toward ownership and freedom might work, because in a way or another we will end up there being technically the sole solution.
The big tech model where trust is in the company, not the person. Business love the big tech model because it's easier to let a few credit card companies deal with the trust issue than establish a trust relationship with everyone directly (or deal with cash), because surveillance capitalism is more profitable, and because it's more profitable to rent than to sell.
The big tech model can profit first on that cost difference, and later on switching costs which would otherwise inhibit abuse.
It has essentially nothing to do with the internet, as mainframes were networked long before personal computers. Even back in the 1980s, POS terminals used dial-up to verify credit card transactions.
The trust problem is easy to solve, with an open society: as long as payments got processed with open APIs and the government takes care of the frauds there is no trust problem. I do not need to trust a third party with eCash, I only need to trust my State protections.
The idea is already tempted, see not only the historic eCash, witch are the modern GNU Taler chosen (it seems) by the EU for the digital Euro https://www.ngi.eu/ngi-projects/ngi-taler/ and https://social.network.europa.eu/@EC_NGI/111499172838284606 but also https://openfisca.org and https://github.com/CatalaLang/catala or few others alike.
That's still embrional but in FLOSS terms we have already more than enough, we just miss the law enforcing it and the schools teaching it to the masses.
As I was on the phone and going through their documents, Windows 10 decided to install updates. I'd experienced this before and had done everything I could to try and configure Windows 10 to require my permission to run updates, but it doesn't work that way at least when you are a small business without an I.T. team.
After a few minutes I told the customer I would call them back when my computer completed its updates. The update ended up taking over 40 minutes to complete. What really bothered me the most is that Microsoft is setting the priorities of our organization - software update instead of resolving a critical customer issue.
I've never had a Linux update require so much time and definitely I've never been spontaneously and without requesting my permission locked out of my computer so Linux could run an update.
"Big Tech", as discussed in the article, appears to me to be no longer concerned with small customers and operating in such a way as to assume we are all just their guaranteed customers so they are free to do with us as they please.
Open source is about licensing, big tech is about scale.
When I sit down at my mac, I have a working and very polished calendar, mail client, todo list, contacts, note taking app, music player, browser, photo editing and library management tools, video call and conferencing software etc. And all of it syncs with my phone and my tablet out of the box.
When I sit down at a Linux machine, I have a calendar that breaks every 5 minutes and I can't share anything with anyone without futzing with iCal feeds and hiring another provider, a mail client that is ugly as sin and doesn't integrate with the calendaring or contact management stuff at all, a job and a half to find a note taking app that actually works properly, a todo list app that syncs with nothing, a spreadsheet package that crashes whenever I try and print something and oh hell I give up by then. And the answer to this? Roll out nextcloud on a VPS. Kill me, with a spoon. This is not freedom, it's just slavery of another kind.
I just want to get shit done. Big tech covers that. Please take this as a recommendation to tidy up all this hell and just help people to get shit done and then it will be an alternative to big tech.
Now that I'm working on a proprietary version[2] (with a block editor I rewrote from scratch), I'm talking to these end users and understand their frustration in using my product. For example, many users had issues discovering the different features of the app, so I created a toolbar, which much helped. This is just one example.
On paper this sounds really good but there's a lot of overhead when it comes to maintenance. "Yeah, it's just one more docker-compose.yml, big whoop"(yes kubernetes is pointless overkill if you are the sole user). I've said that too many times and it's not true cause it only takes one small thing that you overlooked and you have to spend a day or two to put everything back up together.
Another thing worth mentioning is that open source can be a good alternative but open source does not mean free or cheap. For instance, I've gotten really into drones and radio communications lately. Take hackrf and the baby brother that is flipper zero - they are both completely open source but neither of them is cheap. In fact, they are really expensive - they are effectively open source ASIC's. I'm willing to bet that north of 80% of the cost is down to the software and not the hardware - because polishing a piece of software to the point where you can pick up a product and use it without effort or a steep learning curve, involves a ton of work on behalf of developers and UX/I people.
And you can't really cut off all big tech - open source phones are BAD, you don't really have a good alternative to google maps and waze, you still heavily rely on search engines and a few dozen services if you start digging deeper. There are also a number of services which do not have an even half-decent open source alternative. Also not everyone has the skills to set up and run these things.
I think the big case in favor of self-hosting whatever you can is that while open source is far from immune to leaks, if it resides in your private network(which it should) without access to the rest of the world, those holes will eventually be patched and you can take action in the meantime - stop the service, block a few ports, etc. The odds of you personally getting affected are pretty low. Now if a leak happens in big tech, there's nothing you can do about it and by the time you learn about it, it's often too late. Honestly, this is the number one reason I'm doing this to myself.
Imagine if you had to compete producing widgets in a market landscape where some hyper-conglomerate would source and distribute all power, define and install all plug standards and, in addition, produce and rent any widgets that saw consumer traction. For decades this is what has come to pass as normal in this domain.
Openness (of varying degrees), standards-adherence, interoperability and competitive markets are connected attributes. In this context open source is an extreme productivity multiplier. Maybe the most potent such development in modern human history. Entities that adopt open source would collectively out-compete in innovation and usefulness any proprietary offering. But for this mechanism of sharing knowledge to thrive and reach its full potential there has to be a real market for digital technology.
Big tech can totally sell "FOSS services" and provide ground works for it, like some of it do - you don't have to lock people with proprietary stuff.
Even more, big IT tech couldn't exist without FOSS in this form and shape, while the opposite is not true.
Linux only took off during the dotcom days as IBM, Oracle and Compaq started adopting it into commercial workloads, back in 2000.
Visual Studio Code isn't in the same ballpark as Visual Studio. It was already an Azure project, as the Monaco editor, and it was a way to kill Atom.
ARM is only successful on mobile devices and Apple hardware.
If you mean ARM on server, the most successful company, Ampere, is largely owned by Oracle, and there are some ongoing discussions about a full acquisition.
Your "only" is funny. That is by far the biggest computing market worldwide.
How has Microsoft ditched VS for VSCode? VS is lightyears ahead in features and performance.
The two are not even remotely comparable. VSCode is a text editor that wants to be an IDE, but if you work with C++ or .NET you're shooting yourself in the foot if you use VSCode.
VSCode is not a serious alternative to VS or other IDE's like JetBrains Rider.
GitHub is not open source.
Without that leak we would not have the ecosystem evolving around Llama.
This isn't charity, they are literally using more OSS software than they produce their own software. By several orders of magnitude in most cases. Companies like Google have many millions of lines of code in proprietary in house code. But they depend on an even larger amount of code in OSS form.
E.g. Android and Chrome OS are based on Linux. Those products are built on many thousands of open source packages. And of course Google is contributing to lots of them and created a few themselves. Chrome is open source because webkit was open source because Apple forked KHTML from the KDE project.
Open source without commercial companies contributing would be much more of a fringe thing.
VC funded OSS companies are a bit more challenging. These companies are perpetually confused about their licensing and need to keep things proprietary and open at the same time. These projects get a lot of attention because of the VC money but technically they only represent a tiny fraction of the OSS community.
I don't think this is actually true:
1. The Google codebase is on the order of billions of lines of code, not millions.
2. It's basically all written in house, from the threading libraries and core standard libraries up. The parts that are open source (e.g. Linux, OpenJDK) are very small compared to the code they've written themselves.
ChromeOS and Android are open source, but they aren't even close to being the bulk of their codebase.
If Linux had never existed they'd have found some alternative, probably either a bulk licensing deal with a proprietary UNIX vendor or they'd have used Windows as the closest cheap Intel based alternative. Then they'd have put funding into developing their own in-house serving OS a lot earlier.
Source: I worked there.
> Chrome is open source because webkit was open source because Apple forked KHTML from the KDE project.
Chrome is open source for strategic reasons and because the executives in charge wanted it to be. There's no particular reason it has to be open. Safari and Edge aren't.
My conjecture is that open source is polished enough for most customers to use when there are commercial interests implied. Linux on the server is a resounding success, Linux desktop not so much.
People have to put food on their table and can't work for free. Someone has to pay for that work. Nobody will pay for it if he can't extract some benefits from doing so.
Because it is risky. The more esoteric the knowledge gets, the further it moves away from your core business, the more in-demand the skills are. As an example, maintaining your own metrics and timeseries storage. It takes quite a few skilled hands to do this in house and probably only feasible for larger companies anyway. Or you can simply hand this problem over to DataDog. While they are pricey, it is potentially pricier to build your own internal DataDog-like system, especially if you consider the opportunity cost of pulling your most skilled engineers to build it instead of building your product that your customers are paying for. Companies are perfectly willing to pay a premium to not worry about something, and that includes not worrying about your very skilled engineers leaving and then needing to scramble because no one else understands what has been built.
If you need more anxiety, just think about the hottest technology right now that is capable of relating massive amounts of data instantly :-)
SQL isn't ready for AI.
IBM RedHat is the steward of RHEL, and Canonical is of Ubuntu, so if you want support from them there are no real other options, but they do work with multiple different ISVs.
If you want to stay 'independent' and have the most leverage you can take a linux which is not from a bigco, such as Debian.
I'm really worried about cloud-lock-in for bigger companies. My previous company switched large amounts of product to AWS, when I asked about how this was feasible after doing back-of-the-envelope calculations they said: well you should not consider list price, nobody is paying list price, we get discounts.
This reminded me of an ISP I worked for in 2006 that invested in large amounts of Solaris machines because they got big discounts instead of going for the much more obvious Linux route. Then after two years or so the new (Oracle at that time?? I'm not sure) sales rep paid a visit and they said, when they were not able to sell MORE new servers: OK screw the discounts, from now you're paying list price. So that got them stuck in a real bad place. I'm afraid the same might happen to companies who move to cloud providers as well.
And then I've not even touched on issues such as privacy, security, business continuity, and losing the skill to actually run your own hardware
i see a trend in recent years in losing the skill for making proper software too - Something that works and is usable and is not about changing colors and pixels and protocols everyday for the sake of it. Maybe the opposite trend of the everybody-can-program since 1995? Or maybe not - not sure how ML-generated stuff will impact all this.
btw there was ~2021 talk/article by same guy about how outsourcing-everything leads to total deskilling:
[1]:https://www.suse.com/c/embrace-linux-diversity-simplified-mu...
That is forgetting why your enterprise relies on an Adobe service. It is because nobody else has software that does their job as well as Adobe does.
This discussion is non-sensical to me. Of course the license matters.
I'll admit I've never used K8s outside of work very much, so I can't really argue on that!
> What were you thinking of things to take into account for k8s over swarm? The main difference for me is k8s needs a hell of a lot more boilerplate yaml for doing basically anything.
I think that's a big one, yes. Stateful services (i.e. volumes) are also much easier to setup and understand with Docker Swarm - which is the same as Compose. The routing mesh[1] is also lovely. I didn't use the Kubernetes routing mesh at work because the infrastructure department didn't allow us to, why is one a reason I was arguing against it; we used a very powerful and complex system without profiting from one of its most powerful feature.
If anything, the x86 world was more open and more compatible. We have enterprise distros running on both Intel and AMD, supported by their hardware makers. Who in their right mind runs 3rd party Linux distributions on smartphones in production environments (i.e. the CEO's smartphone)?
1. You aren't average. Market forces might not align with your use case. The pricing model might change, or it might happen that you find out later it doesn't scale well for your business.
2. They might leave you suddenly. For example, all the google "products" (quotes because there is actually only one--ads) that have disappeared over the years. Even when an open source dependency suffers a cataclysmic licensing event, you can still fork it and carry on, provided you've both chosen your dependencies wisely and hired the people capable of maintaining them.
3. By choosing a vendor you're making a commitment to ossifying a part of your stack. The observability example is a good one here. At the companies I've worked for who do all their logging, metrics, alerting, etc in-house, developers aren't afraid to use the tools. The tools adapt to the requirements, whether it's cost efficiency, features, whatever. At the companies where we've used vendors everyone's perpetually afraid of increasing the monthly bill, and nobody has a say in deciding what goes on the product roadmap. To be clear, this might be the right trade.
I remember visiting early Hotmail and their sharded “capital unit” was a great big Sun storage server and a bunch of Intel desktop towers running BSD. The latter was considered rather wild at the time.
Last I checked, some eBay URLs still had “ebayISAPI.dll” in them, which is a remnant of that period.
https://www.joelonsoftware.com/2002/06/12/strategy-letter-v/
I was using VM systems running on PCS from 1989 (OS/2) Linux only started in 1991 and did not take off for say 10 years, by then Windows NT existed.
So VM was necessary for Linux but was not the reason for it taking off.
In my experience Linux came in for servers replacing other Unix servers. Windows NT servers continued for some time.
As for desktop you still need Excel and to a lesser extent Word and these are still best on Windows.
Had Microsoft known better, and Linux would never taken off on PC.
Paying for software was never an issue back then, piracy was quite common, you could get whatever you wanted on the countries where street bazaars are a common thing.
Check the list, make your order, come around the following week.
I think you are grossly underestimating how dependent both companies are on various open source projects. It's definitely true that they also do a lot of in house code of course; and they also contribute a lot of their own projects. And of course especially Google is a repeat offender when it comes to creating a lot of dead projects, reinventing the wheel, etc. It looks like their attempt to create their own operating system kernel is slowly dying now. Fuchsia is all but dead at this point. So they are back to Linux being their only future. There's the whole Kotlin ecosystem, which they helped create, which is starting to compete with flutter. And so on.
I'm not really underestimating anything. I worked on the Google codebase for years. It has very little dependence on open source code relative to its overall size.
> what sells in enterprise is marketing and a good sales team, not "effectiveness"
I work for a SaaS company. Our marketing and product teams work hard to convince potential customers that our software fulfills their needs at a reasonable cost.
When we buy services and software we don't look at any of the marketing materials, there are very trough analysis of costs/benefits being made, dollars and even cents are counted. Every cost that can be cut while we can still deliver something to our customers will be cut.
Outside of tech, that's not how it works. They don't have in-house developers. They don't read HN. When they need a database, they ring up a large vendor and spend hundred of thousands a year for Oracle, when a PostgreSQL container would do just fine. They often don't even care they could pay zero, simply because PostgreSQL doesn't come with a phone number to call when the DB crashes.
Paying someone else to solve your problems can also carry huge risks of all sorts. Sometimes investing 1000 dollars into a risk-free solution is better than 5 dollars a month with gigantic strings attached.
Firstly, of course, most customers don't have any programming skills at all, so the point is moot. But let's limit ourselves to customers that do have IT departments.
It's worth noting that IT departments are already busy. Deploying resources because you found a bug in PostgreSQL is unlikely.
But ok, you found a bug. And we happen to have a highly paid C programmer on staff. Let's ask him to take a look.
He's not familiar with PostgreSQL architecture- do it'll take a few days to download the source, make a build, hope the build is close to our version, and deploy to production. (This has already cost the business more than a enterprise support contract from Postgres but whatever...)
He then spends a month working through various subsystems to determine the exact flow that leads to your bug. (I'm gonna ignore all the cases where the "bug" isn't a bug.) He makes a tweak, merges it into the current build, and deploys.
He even submits the PR which may or may not be accepted. Until it is he's regularly pulled from his task to update the PostgreSQL source and rebuild.
Everyone else plagues him either PostgreSQL questions twice a week now. His manager gives him a 'poor' rating at the next review because the thing he was actually hired to do is not getting done.
Unless the company is selling to other PostgreSQL developers, there's no "free publicity". That's not how publicity works.
So yes, your scenario is possible, but its simply not how things happen. You complain to the IT department - they add PostgreSQL enterprise support to the next budget. They're not looking to take on extra load unnecessarily.
Yes, the major contributers to big OSS projects are tech companies contributing full time programmers. And that's a good thing. But customers do not have the time or resources (or inclination) to go down this road.
Frankly, it's too hard (in most cases) to just build the software, much less understand the code to make changes.
If we're worried about the supplier now then we don't bu from them. If they suddenly change down the road (and most established ones don't) then the fallback is either "we don't need support anymore" or (more likely) we start looking for another system.
I've been on the other side of this. We sell a product into an established mature domain. We're the "newbie" on the block. Most of the sales we get in this space are from customers who are unhappy with their supplier. We offer better sales and service, and obviously a smooth transition from their existing data (which we import.)
Neither product is OSS - but even if it was that would be irrelevant to the user. (It would however have made our integration code a lot easier to write, so there is that...) A lot of the users we convert have "bought" their software. It's costing them nothing to use it. But they switch to us anyway (we have a subscription model) because our model can afford to fund full time support staff, whereas the sales model cannot.
So, I think this "choose another servicer" is more of a theoretical than practical feature for most OSS systems. Obviously there's really good support for the really big projects, but basically nothing from 99% of them...
The centralized trust model does not require mainframes connected-to by dumb terminals. We need only look at how Visa in its first few decades used carbon copy devices and signatures, along with eventual consistency across a network of mainframes, to gain market power.
"The trust problem is easy to solve" is laughable, as you well understand by the need for "the law enforcing it and the schools teaching it to the masses."
> Whilst Visual Studio Code is "open-source" (as per the OSD) the value-add which transforms the editor into anything of value ("what people actually refer to when they talk about using VSCode") is far from open and full of intentionally designed minefields that often makes using Visual Studio Code in any other way than what Microsoft desires legally risky...
That is the thing, what really won on server and embedded was UNIX/POSIX, and while GNU/Linux is the cheapest and more flexible way to achieve that, it isn't the only one, and the best experience is anyway with vendor custom distributions with their special sauce, not the pure FOSS one.
I think one can attribute many things to the success of Linux (incl. POSIX) but it's not about one single thing and the whole shouldn't be discounted.
It is no different than using managed Kubernetes, or doing everything from scratch instead.
And I wouldn't agree Linux desktop is unsuccessful. It's actually growing quite a bit in the last few years. And of course ChromeOS is also Linux based and capable of running Linux software. Likewise, MS bundles Linux with Windows and it is widely used by developers using Windows that way.
But even without that Linux Market share is now 4.5%. Quite a few gamers are discovering Linux works pretty well lately. With ChromeOS included it's closer to 6-7%. Linux on the Desktop is bigger than ChromeOS. I'd say it is getting better.
I guess what ticks me off a bit is the trope of dissing on management for saying 'no' but leaving out all the relevant context that might show management to be right.
Don't get me wrong, I've seen my own fair share of bad management. But it's not so black and white when it comes to grand sweeping decisions like "should we invest in PostgreSQL or keep paying Oracle big license fees?"
Why? So we'll have someone to blame when things inevitably go wrong? In my experience, the people who like to say the buck stops with them are nowhere to be found when that happens.
I much prefer organizations focussed on actually getting things right, instead of worrying about who takes the blame.
Say, if the org runs Postgres in-house, there's a mighty chance that an intern somewhere might decide to ...test things out in a creative way.
Perhaps the idea of outsourcing that to Oracle is that Oracle has the processes/controls to rein in such interns. As opposed to e.g. a hospital having to create such processes/controls.
(Oracle is still a bad idea IMHO, just slightly less so comparatively)
This sounds just like a general argument against prenups and specifying any kind of contract penalties?
In any case: one way a supplier can make me less worried about them now is by open sourcing.
I mostly agree with most of your points.
There are aspects of a State to function that MUST be ONLY fully public. Again it's simple, in conceptual terms, hard to get applied in reality.
NeXTSTEP was 4.3BSD plus Mach, using Display Postscript as its windowing system and TIFF as its image format, supporting transparency for icons. macOS is FreeBSD plus Mach, using Display PDF as its windowing format and PNG as its image format, supporting transparency for icons. Basically NeXTSTEP but every component upgraded to its then-modern equivalent. (Except Objective C, they kept that.)
> Bell Labs did the same with Unix, but Unix was still not open source. This is why we run GNU/Linux today, not Unix™.
I know pretty well how NeXTSTEP used to be, my graduation project was to port a visualization framework from NeXTSTEP/Objective-C to Windows/C++.
The fact that UNIX™ later went on to become a compatibility specification, not a specific implementation, is irrelevant to the analogy.