We keep track of EOL dates in our internal wiki with links to various websites. Would be nice to reduce the scatter and just link to this site!
Would you accept Pull Requests that add server hardware EOS and EOL dates? They are especially hard to find and it would be great to have them in there as well!
We track .NET framework separately (https://endoflife.date/dotnetfx), which mentions that the "support-tied-to-windows" policy applies in specific cases:
> On operating systems prior to Windows 10 version 1809 and Windows Server 2019, .NET 3.5 SP1 assumes the same lifecycle policy as the underlying OS on which it is installed.
Thanks for the site. It answered a question I had in the back of my mind, but hadn't found the time to look up.
One suggestion I hope you can look into: the Java world is a lot bigger than just Oracle Java, with multiple implementations where End-of-Life greatly differs. If possible, people avoid Oracle Java. Look up OpenJDK, Eclipse Temurin, Azul, IBM Semeru. There are some more, but I have never used them.
> One suggestion I hope you can look into: the Java world is a lot bigger than just Oracle Java, with multiple implementations where End-of-Life greatly differs. If possible, people avoid Oracle Java. Look up OpenJDK, Eclipse Temurin, Azul, IBM Semeru
Yes, though it would be more helpful to have a comparison chart of the various (Open)JDK distributions, because for most purposes they are exchangeable, apart from the support terms.
Do you have plans to monetize it? I hope not. Because this is a thing that should exist in the world. I would much rather support this via donations than ads or anything else.
OpenCollective says that the total donations are $1,107.75 USD with 6 contributors. I threw another 20 bucks at them after reading that, but it seems like at least ads or something would make sense purely from a financial perspective, since even with 8M impressions, not that many are interested in donating sadly.
I do really enjoy the site, it has lots of technologies listed and looks very pleasant. Actually used it just this week, since I started building an API with the previous .NET LTS version but a new one came out, so I suddenly started having to install older package versions for compatibility reasons hah.
> OpenCollective says that the total donations are $1,107.75 USD with 6 contributors. I threw another 20 bucks at them after reading that, but it seems like at least ads or something would make sense purely from a financial perspective, since even with 8M impressions, not that many are interested in donating sadly.
What are the operational costs of running such a site?
Currently none, but that's mostly because we've worked hard to keep it that way. I publish finance updates on a GitHub Discussion (https://github.com/endoflife-date/endoflife.date/discussions...). We used to pay $9/mo for Netlify Analytics but didn't find it helpful.
We also have a pretty long roadmap (https://github.com/endoflife-date/endoflife.date/issues/2108) that focuses on integrating better with the sbom ecosystem (We want to make an API that does SBOM->EOL alerts for eg). Such projects require a lot of development effort, and we'd like to be able to sponsor it on our own (via grants/donations).
> We want to make an API that does SBOM->EOL alerts for eg
You could consider making more expensive features paid options? Definitely tradeoffs there but it is an option, especially for business-useful features
Setting aside a HN spike, $100 a year should be fine for hosting. Or use someone else’s hosting platform like GitHub, cloudflare etc for the cost of a domain name.
Could be a very useful resource. There have been a couple of occasions when I've found it harder than it should be to find EOL information.
Of course sometimes there isn't any because many small projects have no official support cycle, but sometimes projects of significance, and even some commercial products, don't either. But sometimes there is and it isn't well documented (or sufficiently linked so it is hard to find).
Would you consider adding a timeline-like view, for instance pages listing everything due to go EOL or move between support categories (current, still under standard support, extended (full LTS), extended (limited, i.e. security only or common packages only)) this/next month/quarter/year? Those pages could be static and refreshed daily rather than needing resource to generate on each visit (filtering could be done client-side to allow more flexibility without extra server resource for dynamic generation). This could be done by a 3rd-party, but that might involve scraping the API/site in a way that imparts an unfriendly amount of load (unless I'm missing something, the data in the release-data repo doesn't contain everything that would be needed for this).
Yes, we have an open issue for such timelines (https://github.com/endoflife-date/endoflife.date/issues/2148). Your comment has given me the idea to implement it on our "tag" pages, so you could see the timeline for all java-runtimes in one place for eg.
Generating pages isn't hard for us. We maintain it in our frontmatter/yaml (See https://github.com/endoflife-date/endoflife.date/blob/master... for example). The release-data repo is for tracking new releases of upstream products so we can patch the latest version automatically, and get notified of new major releases that are missing in our tables. release-data is a much smaller subset of our data, and doesn't include the critical EOL data.
In the case of commercial products it is especially frustrating if you need to log in to some customer portal in order to find such basic product information. And even if you create an account and log in it is not guaranteed to find this information.
I can relate to a certain extend as some companies might not even know when they plan to end support for a product, but even this is information that is worth sharing with their customers.
I think calling it the Hall of Fame/Shame is unjustified.
I certainly agree when we talk about hardware products, because having very short support periods for those just seems wasteful.
But it gets difficult when entering software, because it probably differs from user to user how fast they would like the software to evolve.
I can think of software that I want to be boring and work forever the way it does now without changing (except security bug fixes ..). So support periods are long and consist only of patch releases.
And then there is software where I am waiting eagerly for new features to be added and would like the maintainers/the company to focus their attention on it if possible, thus taking resources from the maintenance side of things and moving them to the feature-factory. So support periods may be short and major/minor release bumps happen frequently.
Having one of these products in the Hall of Fame and another rot in the Hall of Shame does not seem to add any informational value.
I'm not sure if that's a good idea. Maintaining long-term-support, especially for open-source projects with volunteers is a unacknowledged (and often unpaid) work.
It might make sense for certain categories (Mobiles, tablets) but it is so hard to get good data for those.
I just want to say that I love this and have it bookmarked! I've moved into my 30s and am now thinking about life "maintenance first" and this will be a helpful resource for planning.
I imagine the only help y'all need is for us to update EOL info when we see it?
Yes, great it like a Wiki. Additions of new products are also welcome, as is building tooling on our API or improving our release automation so we track releases better.
Could there also be something like a comparison view of devices and models where you can see overlapped timelines when support ended (e.g. in the past vs now?)
Also, something like repairability scores from e.g. ifixit would be amazing to integrate.
Whenever I am looking for devices, I am also on the lookout for whether or not a third party OS has upstream support for it at the time of buying (e.g. LineageOS or postmarketOS wiki) because the device isn't "just a dead thing" when there's an OS available that you can flash on it.
It turned out that it is too hard to generate clear charts with vague data. We often only know whether is device is supported or not (true/false, see comments about samsung below in this thread), and don't have clear release dates.
I'll get to it someday (PRs welcome), but it might not work for the usecase we want (picking phones) because data on mobiles is very vague.
repairability score -> sounds interesting, will file an issue and see. The hard part is that there's no clear identifiers for devices (SWID/CPE are just not good enough) for us to track this kind of data from elsewhere easily.
Great effort! As I’ve realized the folly of updating my phone every 2-3 years, have you considered a sign up to be notified that my device is reaching its end of life?
As an example, my cellular provider gives a discount assuming I’ll be BYOD for the next three years. I’m really curious if I can make it past those three years. It would mean my phone is 5 years old at that point.
Stock RSS feeds don’t work for future events, so they aren’t perfect for our usecase (warn users in advance of an upcoming date), so we offer ICS feeds instead.
RSS feed of just date changes might work, but ‘s hard to differentiate between a change that creates a feed notification vs one that doesn’t (new release, typo fix, date change by 2 days, and so on…)
- Would like a time line view: as of the next few months, what are expiring.
- Would like to see a stability score for a Product / Stack, you define the metrics for what is considered stable vs obsolete. How it compares with other similar categories.
Second is very subjective. If there's onething that I've learnt over the years working on this - it is that the definition of "support" is very fuzzy. You can't even compare what "LTS" means between different stacks/products, so it is hard to objectively rate anything. It still might be worth doing some analysis and publishing as a blog post though, but i'm not sure if it can be automatically calculated.
We keep track of EOL dates in our internal wiki with links to various websites. Would be nice to reduce the scatter and just link to this site!
Would you accept Pull Requests that add server hardware EOS and EOL dates? They are especially hard to find and it would be great to have them in there as well!
Edit: I should have taken a look at open issues first, it seems there is tracking issue for network equipment so I guess adding server hardware is certainly not off the table (https://github.com/endoflife-date/endoflife.date/issues/1387).