MACH Alliance Interoperability Standards
Apparently, within hours of me publishing a somewhat scathing LinkedIn article in which I expressed my lack of perceived value from the MACH Alliance:
?This appeared:
Let’s dive in. Right off the bat, I was curious when this appeared. View source seems to indicate content from Contentstack, with dates that may indicate its generation within the last month or so. But that is speculative, and I’m not going to ask. At least they picked a good CMS to manage the content, even if that could show bias towards a particular vendor. And if this was a commerce site, I think we all know what vendor it would use.
At first, I was embarrassed that I had overlooked this resource. Then I was super excited! The MACH Alliance had applied domain-driven design to define industry standards for API endpoints, HTTP headers, query string parameters, and JSON payloads for at least content management, commerce, and search! That would be a huge task, requiring vendors to buy into something that actually works against their interests, and could take years to evolve and implement. Vendors would likely (though begrudgingly, at least internally) wrap their existing service APIs to implement least common denominator functionality conforming to the MACH Alliance standards at a new endpoint. This would be verifiable!
An immediate concern was that the MACH Alliance would license these standards, which would likely work against their intentions. Even in that case, all of this would still be in the best interest of the customers.
领英推荐
Unfortunately, I don’t find any actionable content. It’s like that they replicated their marketing site in markdown at github. Seriously, there is even a PDF whitepaper. And a diagram generator.
Possibly the most valuable resource in this repository is a reference architecture diagram, but that appears quite oversimplified and isn’t necessarily an architecture that I would promote. Apparently, by standards, the MACH Alliance means reference solution architecture.
I do appreciate that the diagram includes layers for digital experience composition at the front-end (which I would say is optional – users want to compose in the CMS, not a separate tool), and more importantly, brokerage around orchestration and integration. Again though, there is no guidance on how to achieve these objectives. Prospects are left in the dark, potentially after actually spending time reading the content.
I have trouble squaring any of this with the statement on the website, “Interoperability is a core focus for the MACH Alliance.” By this I assumed they would promote the ability to easily replace SaaS applications in a customer solution. Instead, it seems that they’re just continuing to hammer on the potential value of service orientation. Are there any technologists in the house?
Anyway, maybe I missed something, but none of this content seems in any way specific to MACH architecture from a technical perspective. Maybe ChatGPT can figure it out. I’m not saying that there’s no valuable content here, just that it’s not valuable to me. I kindof just gave up after a few minutes. Hopefully this evolves into something useful, but I have my doubts.
CEO and Founder at Ucommerce: Composable Commerce for .NET Developers
5 个月You caught my eye with this one. I was SO hoping to see actual interoperability standards. Alas it was not to be.
Great! Your insights are why we started the interoperability initiative in the first place. It's frustrating when value is not delivered due to misinterpretation or poor implementation. I have seen mu share of "MACHoliths"... We all want to avoid those. Many businesses benefit from composable architecture where it makes sense. None of my clients have 100% composable ecosystem. They typically compose where differentiation is high, multiple platforms with the same responsibility (eg CMS or ERP) need to coexist, to isolate legacy systems, or enhance capabilities like pricing logic and customer engagement. The interoperability standards effort aims to guide integration strategies based on SI and end-user experience. Whether to compose, choose out-of-the-box, or use a mixed approach depends on your business model complexity and solution lifetime - I am sure you agree. Interoperability is a step towards achieving the flexibility that composable architectures promise. Your feedback is appreciated. Specific examples and case studies are next. Want to contribute? :) Let me know. Adam – Interoperability Working Group More depth to the ref arch: https://github.com/machalliance/standards/blob/main/reference/architecture-explained.md
Coaching SIs, agencies, and customers to improve success metrics for digital progression and experience management initiatives. Blogging about composable solution architecture at deliverystack.net.
5 个月I'm full of crazy ideas. What if you could pass zero or more transformers with the JSON of every API call, little fragments of JavaScript or whatever. These would convert your normalized JSON to the vendor's proprietary API requirements including URL paths, HTTP headers, and query string parameters. And transform their JSON on the way back. Then, you'd only have to change transformers to replace systems. Is that possible? bad idea? Would vendors go for it?
VP Product at Crunchr
5 个月Alex de Groot I remember those meetings. And I’ve since been involved in similar meetings in other domains. My feeling on this is that you won’t get to interoperability if you can’t agree a standard to push to. And after much soul searching I think that, because this is all in service of solving for a use case, you won’t be able to come to one agreed standard for the output because it depends on what you want to use the data for. And there is also always the fear of missing something and the tendency to want everything. Soon you go down the path of a very generic model that could be used to describe anything and ends up describing nothing uniformly. So we are back to where we started, no easy interoperability. And of course, we haven’t even started discussing the commercial concerns that would hold vendors back. Will AI eventually be able to solve some of this? Not while it’s still limited to garbage in garbage out. So for now, i’d like to postulate that interoperability is an ideal and the best solutions automate as much as possible and make it easier to handle the rest.