Donate

[Qqmike]

Just want to plug and encourage donations both to Kubuntu forums here and to Wikipedia. Both good places to allocate some of your donation funds, however small. I don't mind saying that my Wikipedia donation is only $10 ("once only," but have done so 3 years so far), and they really appreciate it. And of course donations here to Kubuntu forums is always appreciated. (kubuntuforums.net: Go to Settings > My Account > Paid Subscriptions)


I know one of our mods would prefer not to solicit donations for Kubuntu forums so explicitly, so just let me say that my post is on my own, my opinions only. --Mike

[Don B. Cilly]

Uhm.
I don't know about KFN. But have you looked at Wikipedia's audited financial statements? Take the 2017-18 one.
The lore they're selling is "a few underpaid volunteers working out of mothers' basements". And "please donate, we can't afford no shoes".
Now...
- Net assets at beginning of year: $113,330,197
- Net assets at end of year: $134,949,570
- Increase in net assets: $21,619,373

Ahem. Some expenses:
- Awards and grants: $13,555,339
- Travel and conferences: $2,389,279
- Salaries and wages: 38,597,407 (underpaid scrawny kids indeed).
And...
Internet hosting: $2,342,130.

Now, 2,3 million in hosting?
Now, we all know (well, maybe not, but it's an interesting fact) that the whole of Wikipedia, images included, sits comfortably on a pen-drive.
Sure, the bandwidth. Er, well. A T3 line costs less than $40,000 a year.
How many would you need for Wikipedia? My guess is, one. Maybe three, if you want to really overdo it.

Now, I dislike being suspicious, but sceptical I have to be.
In a world where "nonprofit" is increasingly becoming synonymous with "fraudulent", the first thing I do when asked for a donation is look at the financials.
And I may be wrong, but those look slightly fishy to me.

[Edit] Boy, the US$ is high at the moment :·]

[jglen490]

So I guess you're saying, in effect, "Screw those pirates, you're a fool to contribute"?

Yes, some non-profits seem to be more efficient than others. Some ARE downright crooked. But statistical and financial "analysis" is in the eye of the beholder. I've contributed to wikipedia, along with Red Cross, Goodwill, Salvation Army, Lutheran Disaster Relief, and a host of others across the years. A lot of Grumpy Guses say "DON'T GIVE TO THOSE GUYS, give to mine instead". Reach into your pocket, give what you can, to whomever YOU want to give, and press on with a clear conscience.

[Qqmike]

Having been a professor of business and a business consultant, I'm very skeptical of financials! and skeptical of HOW accountants actually count the beans. If you're questioning someone's financials, bring your questions to their attention to get their response.

[whatthefunk]

Now, 2,3 million in hosting?
Now, we all know (well, maybe not, but it's an interesting fact) that the whole of Wikipedia, images included, sits comfortably on a pen-drive.
Sure, the bandwidth. Er, well. A T3 line costs less than $40,000 a year.
How many would you need for Wikipedia? My guess is, one. Maybe three, if you want to really overdo it.

This seems reasonable to me. Its not just about bandwidth or about a single T3 line, its about availability. I work in IT at a web company. Our small company has a hosting bill of around 2,000 US a month, and this is going to be increasing so that we can increase redundancy. Check here: https://meta.wikimedia.org/wiki/Wikimedia_servers
- Five data centers
- Several DB clusters (each node in each cluster is probably on its own server)
- Memcached (again probably on its own server, and there are probably a lot more than just one server)
- Elasticsearch (yep, on its own server and likely distributed)

Most IT companies arent just sitting on a server rack in the basement. They are distributed around the world in clusters with multiple layers of redundancy. This is very expensive to build and maintain.

[Don B. Cilly]

Yeah, well, I've seen that Wikimedia servers thing too.
Still, also considering the fact that they get a lot of bandwidth "as donations"... I still find the 2.3m figure, well... ludicrous. For 50 GB total storage. more so.
I still may be wrong...

On other notes, I do contribute to Wikipedia. I write. And the fact that I have "Supporter-Active" written down the side of my profile... *

* In case you're not familiar with the "And the fact that you've got "Replica" written down the side of your guns, and the fact that I've got "Desert Eagle point five O" written on the side of mine, should precipitate...", well, you can look it up

[whatthefunk]

Its not about storage or even bandwidth really, its availability.

Check out their main stats page: https://stats.wikimedia.org/v2/#/all-projects

All of Wikipedia got 20 billion page views last month. 20 billion! There are 2,592,000 seconds in September. That means they got around 7,700 page views per second. Even a site serving a single one byte file would not be able to handle that without a massive amount of infrastructure.

As I said above, our company pays about $2,000 a month for our servers (I made a typo earlier and said 20,000..) We did some stress testing not too long ago and found that around 30 concurrent requests will bring our site down. By concurrent I mean 30 users making requests at the exact same second over an extended period of time. If we got 7,700 page views per second, our infrastructure would completely collapse. The whole thing would go down in a massive ball of flames. Our servers simply can not handle that volume. They don't have the CPU or memory to deal with it.

We use a load balancer to distribute requests across five servers, but that number of requests would probably kill the load balancer before half the requests even made it to the servers. Even if our load balancer and web servers were beefed up, our other servers hosting elasticsearch, cache, database etc would die. So we would have to beef them up to. That alone probably wouldnt be enough so we would have to increase the number of slaves we use, or migrate to a cluster server architecture. You can see that if we had that number of requests, our infrastructure costs would go up exponentially. So 2.3 million sounds fair to me.

[jglen490]

@whatthefunk ... agree, absolutely.

The best enterprise systems consist of multiple iterations (not versions) of themselves.

[MoonRise]

This seems reasonable to me. Its not just about bandwidth or about a single T3 line, its about availability. I work in IT at a web company. Our small company has a hosting bill of around 2,000 US a month, and this is going to be increasing so that we can increase redundancy. Check here: https://meta.wikimedia.org/wiki/Wikimedia_servers
- Five data centers
- Several DB clusters (each node in each cluster is probably on its own server)
- Memcached (again probably on its own server, and there are probably a lot more than just one server)
- Elasticsearch (yep, on its own server and likely distributed)

Most IT companies aren't just sitting on a server rack in the basement. They are distributed around the world in clusters with multiple layers of redundancy. This is very expensive to build and maintain.

+1 to that. Same here on what I handle and know of hosting.

[SpecialEd]

Desert Eagle. The big gun all the little guys bring to the range.

[benny_fletch]

And of course donations here to Kubuntu forums is always appreciated. (kubuntuforums.net: Go to Settings > My Account > Paid Subscriptions)

+1

Donate if you can. Of course, my opinion as well ;-)