I, for one, welcome this change. OSM brings many benefits including the ability to improve the map layer.
That said, route planning in Ride With GPS is better because you can switch between different maps -- and even use Streetview whilst exploring Open Street Map!
Ultimately, Strava is a great tool for segment & performance comparison but routing is streets ahead in Ride With GPS.
Bytemark | York, UK | On-site & Remote | Office support administrator
£15000 to £33000
This role is for someone who loves being on the inside of a business, enjoys supporting both customers and staff, and can manage the logistics of a busy office. All experience levels considered.
Bytemark | York, UK | On-site & Remote | System Administrator
£16500 to £29000
This role is initially centred around customer service and building internal systems. Ultimately, we are looking for "power users" or Linux/Windows server administrators looking to build their skill sets.
It's really a shame when people can't look past the surface, and just emit vague subjective rebukes like this. XML may not be fashionable today, but really it poses no technical issues compared to the kinds of actual hard technical issues involved in making an open and decentralized IM network that XMPP has been solving for 15 years.
Using someone else's wrapper library is like (insert sufficiently disgusting analogy for doing something with a pile of excretia).
At the end of the day, the underlying product is still an overwrought, overly verbose mess, and the wrappers only serve to hide that. And following the law that all abstractions are leaky, you're going to have to touch it eventually.
A better replacement would be IRC. It serves almost all of the XMPP use cases with no extension, and i'd wager all of them with the right daemon and right plugins.
IRC is better and actually more popular. IRC solved the problem of instant messaging over 20 years ago, and solved it very well. This is a problem that everyone and their mother attempts to re-solve and they always end up with a solution that's inferior to IRC. Just deploy an IRC server folks!
IRC solved basic communication, but the necessity of bots for things like authentication for usernames, logging, and retaining permissions are inelegant kludges, in my opinion. If XMPP has problems, so does IRC.
I don't think it'd be very difficult to set up some sort of web UI for account creation, combined with existing in-client features for usernames/passwords. That'd be more worthwhile than building a new solution from scratch.
IRC has advantages over XMPP by having a much wider variety of clients and the fact that you can create your own integrations for things like build automation in a very short period of time thanks to the simple text-based protocol.
Once you realize that everything on an IRC service is tied to a user account, it makes a bit more sense. And extending from there, different user accounts are responsible for different things. (The *serv accounts on most networks).
I always thought it made sense in a way - it makes for a very neat separation of concerns and makes the stack more modular, since you can now use almost any combination of IRC daemon and services package.
I appreciate the separation of concerns, but I meant more that "agents and clients sharing the same channel of control" is inelegant. It can lead to race conditions in netsplits, even, from what I recall!
Bouncers. They aren't an ugly hack, are easy to deploy, and make the experience way better. I'd prefer to host that sort of thing on my own infra, too.
matrix (https://matrix.org/) is a more modern messaging protocol than XMPP; it revolves around HTTP and JSON because it's the most logical today the way XMPP is based on XML because it was the most logical at the time. Let's hope they can achieve something here !
I also wanted to come in here and promote matrix. You can use it internally, or you can federate it. It takes a lot of concepts from XMPP, and is essentially json over http with an "eventually consistent" synchronization between clients and servers.
