Another fun story: For the longest time the Yahoo pipes front page had an example pipe that merged search results from various online sites (amazon/ebay/cl). It was made by a former employee and was easily one of the most popular pipes. One day we found out he had his affiliate id in all of those links. We chuckled and moved on[1]
Yahoo Pipes was awesome and born from the dream of an open web with open protocols and APIs. It was sad that this vision died and we ended up with so many closed "walled gardens".
I hope we see a resurgence of this ideal, the recent expanse and interest in Mastodon, the Fediverse, and ActivityPub give me hope.
Matt Mullenweg has committed to adding ActivityPub to Tumblr [0], if others in the established web ecosystem also make such moves maybe this will happen!
I've started working on a custom ActivityPub implementation because I am interested by it, but I have to say my impressions so far is that it is more complicated to implement than I originally expected. I find the docs lacking, the test suite hasn't worked in a long time, either. Someone even made a document explaining how to read the relevant specs[0]
Right now Mastodon seems to be the biggest player, but curious what the second/third biggest would be? Pleroma? Having Tumblr add compatibility seems like it could be huge.
> ...Right now Mastodon seems to be the biggest player, but curious what the second/third biggest would be?
Between 2008 through about 2016, it was tough to state which was the most widely used (oh, you'll get opinions, just not sure if stuff was tightly tracked)...But after 2016 - when mastodon came onto the fediverse scene - it certainly has been the most popular software to use. As far as which is second...well that depends on whether you count number of nodes on the Fedi, or numbers of users who use a particular software stack. If matrix federated network is excluded from any list, then second might be peertube nodes (https://the-federation.info/#projects). But if counting by number of users, then second might be diaspora (https://fediverse.observer/stats). Of course, the Fedi being what it is, numbers - while likely good - are only approximations...these data only tell a tip of the iceberg. I mean, the fediverse has been around - and active - since around 2008, so you can only imagine there is some small percentage of Fedi instance admins who prefer not to be counted. So, stat sites like i cited will give you an idea, but not exact numbers. ;-)
The sad situation is that most of the devs hurried into building applications that speak ActivityPub instead of creating libraries that can be used by others.
If you're working with go, I created some libraries that are overlayed over the specification pretty closely. You can find them at github.com/go-ap
Yeah I would have expected more libraries to be available for this as well, maybe if I finish my implementation I'd consider following your example and making libraries out of it. But I only have a hobby use-case and historically I never finish those lol
It looks like it only does client to server (or at least their guide only does this). I'm working on server to server. But for me it's a hobby project anyway so I'll just keep trucking along with my stuff.
But I'm also not developing with NodeJS so wouldn't work for me anyway.
During the app gold rush I would use Yahoo pipes to build 'backends' for apps that were nothing more than screen scrapers.
For example, you want to make an Android app that presents a home screen widget to show the current lottery jackpot or winning numbers. You could build a Yahoo Pipes thing to scrape and parse that, and format it into an RSS feed. The best part is anti-scraping measures allowed it because they thought it was a Yahoo crawler.
Any mention of Yahoo Pipes always makes me think of Tom Scott's "This video has x views". Feels quite relevant to other comments people have been making here.
https://www.youtube.com/watch?v=BxV14h0kFs0
Besides Zapier and IFTTT, what modern alternatives are there for Yahoo Pipes? Ideally with similar UI of a graph rather than "If this, then do that" in a sequential list.
Initial version of Pipes was linear. Mid development we realized it was too limiting and decided to introduce “wires” to allow graphs. It meant a much more complicated UI (thanks to JT for pulling it off - mind you this was back in the IE9 days, no standardized canvas, SVG etc).
I actually used Yahoo Pipes at a job back in 2008, the concept of data mashups was really interesting and we tried applying it to mesh our GIS data with other sources to mashup and enrich the results from our databases.
I don't remember much from that project but do remember having fun with Yahoo Pipes interface, I had completely forgotten about it until this popped up.
Yahoo at the time was clearly underestimated from a developer point of view. Everyone just saw them losing ground to Google in web search, but they had a really strong development team at the time. Lots of great stuff was open sourced, even though they did not win the same status as Google
- YUI was a really good frontend framework based on components and events. It was so easy to create complex interfaces compared to others at the time.
- Hadoop open sourced big data management
- Vespa was a really good search engine in terms of performance and functionality. It seems to still be alive at http://vespa.ai
- They provided several useful APIs without the restrictions that Google and others often placed on their products.
They made some bad strategic decisions and other business decisions, but they had a lot of exciting technology that was often overlooked.
Yahoo should have pivoted from "web services" to IaaS like Amazon did with AWS, because it was no secret that Y! employed some of the best data-center folks besides Google and Microsoft. Could have been a different story?
I've used it briefly. I don't necessarily think it was too early but it was such a niche tool, hard to find, barely advertised to its potential target group, lacked documentation, demos and a community. I think, it was simply a result of Yahoo's chaotic product management at that time. They had a few gems that didn't get the attention they deserved - internally and from the public.
I find them quite limited but I understand the need and also that not everyone can parse a structured response. It's nice that they have free versions though.
Although more complicated to use but also free, I get a lot more functionality from PythonAnywhere, Cloudflare Workers or Oracle Cloud free tier.
I was at Yahoo from 2004-2011. Yahoo often had products before the market was ready for them. Sometimes, the product would hang on and grow with the market; sometimes it would get shut down.
In some cases, like mobile, Yahoo was both too early and too late.
There was a good amount of time where Yahoo would launch something, shut it down, then Google would launch the same thing to much success. Of course, now Google is in the habit of shutting things down, another thing Yahoo did first ;)
I used to think there'd be a lot of value in going through Yahoo PR launches from N years ago and if Yahoo had shut it down, consider if it makes sense as an independent business. Although it can be hard to tell why they shut things down, so it might not be obvious which things were the product worked but the market wasn't there and which things had non-obvious issues that made the product unworkable in practice.
At the risk of appearing to take a cheap shot - I think if this was built at Google there’s a good chance it would’ve been abandoned or cancelled there too
Wow, I had no idea! What an awkward name, too. It's funny, "Pipes" doesn't on its own doesn't stand out as a spectacular name, but compared to "Mashup Editor" it's clearly superior. I wonder why - one syllable vs five, with a nod towards *nix pipes maybe?
Both Mashup Editor and the fact the name is based off Unix pipes is called out in the Wikipedia article, as well as describing the difference (albeit trivially) between Google and Yahoo’s offering.
One sounds like an actual product and the other sounds like an add-on to something else.
"Yahoo Pipes! $7.99 a month with a generous free tier!"
"Mashup Editor included with every Google Cloud subscription above $5/month."
You never want a product to sound like an add-on: Add-ons are low-value and generally used to rent seek off of people who are already locked in. Naming something in a way that it feels like its own standalone product will inspire consumer confidence.
Generic word + descriptor is an easy way to sound like an add-on.
"Video Editor" could be for anything, but is certainly not an independent product.
"Pachi Video Editor" sounds a bit more like its own thing.
"Pachi Cutting Room" is closer to the mark, despite being two words and a company name.
Yahoo Pipes was the direct inspiration for jsPlumb, name and all, back in 2009. We were surprised they hadn't open sourced the UI and decided that somebody should make something like that available. Big fan of Pipes.
I started using the newsboat terminal RSS reader recently and it has some built in filters and ability to call out to scripts and it made me think of this.
Really interesting to see this and YQL because so much changed but at the end of the day a lot of problems can be abstracted/solved as data integration problems. Currently working on http://github.com/cloudquery/cloudquery
I miss YQL much more than Yahoo! Pipes. My memory is hazy, but I think YQL is what ultimately killed Pipes actually, then YQL died because everything that didn’t generate revenue at Yahoo! died.
[1] https://news.ycombinator.com/item?id=9661133