The Boring Enterprise Nerdletter #8: Azure, Distributed SQL, SAP Fieldglass, NLP, RPA, TL;DR
Hi there,
As the old joke goes, everyone’s favorite radio station is WIIFM (“What’s in it for me?”). Enterprise technology is not an exception to this. It’s important not to be blinded by flashy sales slides and buzzword bonanza. We need to understand what technology does and what value it adds.
Over the last few years, there seems to be an uptick in occurrences of “tech without a use case”. There is even a special term “shelfware” invented for software that is bought but not actually used.
We will continue doing our part to inform you about the tech that is useful and the one that deserves more WIIFM questions. To prevent more blockchain-like “solutions”, we all should be asking software vendors: what’s in it for us or even “where's the beef?”.
-Jelena and Paul
Give Me Cool Dev Or Give Me DEATH
Tim Zöller has a few issues with how enterprise software development is done, with "Enterprise Software Killed The Developer". Eye-grabbing title? Yes. Rings true? Also yes. I've worked directly and indirectly with enterprise development for 15+ years, and I understand the perspective that enterprise dev work contributes to stagnating skills and motivation.
He goes through several classes of causes, including over-use of standardization and over-layered architectures. A possible end result - which, again, I've seen in real life - is that developers spend too much time working to stay within the curbs of the over-layered architecture and strict guidelines. I've been a part of teams that just...can't actually do anything for all the hand-wringing.
I get it. There's a balance between "maintainable at the expense of anything else" and "trying the coolest new thing just because it's there". Enterprise software devs have to weigh the former more heavily than startup devs. But too often the maintainability mantra leads to a situation where the whole team has out-of-date skills and can't adapt to business requests for new features. There has to be a way for enterprise software to allow for exploration and occasional "let's try thing thing we haven't tried before" that fuels software developers. PM
Microsoft+SAP Ninjas
Recent episode of myNewsWrap podcast by Christian Lechner featured an excellent GitHub repository with a cornucopia of resources for those interested in SAP and Microsoft integration. These days, this category should include pretty much any SAP professional: it’s difficult to find an SAP customer or partner that doesn't use Microsoft’s products, such as MS Office, SharePoint, Power BI or Azure DevOps.
The repository is frequently updated thanks to the efforts of many SAP+Microsoft friendship enthusiasts, whose photos are prominently featured on the same page. Bonus content: in this Twitter list, you will find the profiles and Twitter posts of these cool "ninja cats". JP
Through The Looking Glass At Fieldglass
Stephanie Gary over on the SAP community blogs about SAP's Fieldglass solution for vendor management. She did a great job of nerd-sniping me, because she threw out key words and acronyms that make my brain light up: AI, ML, and RPA.
She starts off great: "What is the best thing about the use of AI, ML and RPA in SAP Fieldglass solutions? It’s that users don’t even know it’s there." This is the key piece of any solution that uses those things: it CANNOT be the central selling point of that solution. A customer is not buying TensorFlow implementations, they're buying invoice consolidation or candidate-matching.
(You could argue that creating a blog post all about how Fieldglass uses those technologies violates the "users don't even know it's there" policy, but I think since it's on a vendor tech blog it can probably slide.)
I am left wondering, like with all products that have a hype machine around the AI/ML/RPA tech: are those features actually better than their non-enhanced counterparts in other products? I don't know how to measure it, but I think maybe the final piece of the puzzle on making AI/ML/RPA do the right thing for your product is to put full, polished details on the benefits. If AI candidate matching reduces costs by 35% versus simple keyword matching, then it's OK for marketing materials to say it. PM
SAP iRPA and Workflow’s Arranged Marriage
While celebrating Valentine’s Day, you might have missed SAP’s announcement about general availability of a new product called SAP Process Automation. As it frequently happens with SAP’s addiction to renaming and repackaging, “new” here is mostly the name. SAP Process Automation (which we are not allowed to abbreviate) is a compilation of existing products: SAP iRPA, SAP Workflow Management, with some related services thrown in for the giggles.
It makes sense to combine iRPA and workflow in a single editor (after all, workflow is one of the few things in SAP that had a graphical editor since before “no-code” became a buzzword). Combining existing products hardly makes it a new one though (and the corresponding Twitter account was created in 2010). And, more importantly, how is "new and improved" product better than competitor’s offerings and what value exactly does it bring to the business? The product’s “get started” page says it’s “simpler” and “faster”, but “faster” is not a number and “simpler” compared to what? Manual process?
SAP’s marketing department is now out in full force with webinars, community calls, and other product evangelism. This seems like a perfect opportunity for customers and partners to ask questions. JP
Distributed SQL, Distributed To You
I just finished a cool project. With a couple of Cockroach Labs' resident smart people Charles Custer and Jim Walker, I wrote a long-form report on the topic of distributed SQL. You can get a (free!) copy here.
(Long-time readers may remember me mentioning Cockroach in our blowout 2021 year-end issue.)
Though it is sponsored by Cockroach Labs, it is emphatically not a Cockroach sales brochure - other players in this space like YugabyteDB and (of course) Google Cloud Spanner get their due. We take the distributed SQL concept through its paces in establishing a "distributed mindset" perspective, briefly dive deep on CAP theorem and the Raft consensus algorithm, and show off some real-world examples of businesses choosing distributed SQL solutions to globe-spanning challenges.
The final line of this publication is the single most fire-emoji end I've ever put out there. Jim, Charles, and I all contributed lots to this report, but I hereby claim that wrap-up bit as my very own.
In a distributed world, you can create something that delights hearts no matter where they beat.
If you're into databases, data, application architecture, or anything super-nerdy, check it out. PM
Too Long; Didn't Read
Everyone likes TLDRs, the short summaries of long texts. And now, thanks to the advancements in AI summarization, everyone can get their own “executive summary” with a click of a button.
SAP space in particular is ripe for TLDRization. When searching for what an SAP solution <insert name> is and what it does, in the official sources we typically find something like “it’s a Hyperautomation Low Code HANA Cloud Digital Business Transformation solution whose mystery is only exceeded by its power”. But a friendly expert can tell you right away “oh, it’s just middleware”.
As an experiment, I’ve copy-pasted random SAP press releases into two free online AI summarizers: Quillbot and Sassbook. The results, as a linkbait post would say, “will surprise you”. Both platforms distilled excellent TLDRs that preserved the essence of the text while reducing all superficial fluff.
Opportunities for AI summarization are endless (this article has some great examples). And it’s not limited to the texts, AI can also summarize videos. This sounds like the dream come true for busy professionals and TLDR fans. JP