4 ms·

In the requirements, OR seems pretty adamant that they want a COTS solution with no or very little custom code. I'm sure they could do this project with OSS pr

by joshAg 12y ago

In the requirements, OR seems pretty adamant that they want a COTS solution with no or very little custom code. I'm sure they could do this project with OSS products and custom code integrating everything, but those don't come with support contracts. If there's no support contract that means OR has to support it itself, which means hiring people with deep knowledge of the system to quite possibly do nothing for long periods of time until a bug is found for their particular use case, because no one else is using the product. That's not cost effective compared to an off the shelf solution from a vendor that can use 1 support staff to help multiple customers, especially when you realize this means requiring at least a security expert, a db expert, and a frontend UI/UX. It's exactly the same reason why tech companies don't write everything themselves from OSS projects. Sure, they do that for certain things that they deem critical, but not for things like salesforce or the direct deposit and accounting software and sometimes not even for stuff as important and tech-ey as a vcs, like perforce. I'll totally agree that OR should keep IT in house for this stuff, but I don't think it's a great idea for a state or business to suddenly get NIH syndrome and want to write custom software themselves for anything their agencies might need. Can you imagine if this had happened because OR did this project in house and had the same results? Or if the whitehouse had said they weren't outsourcing healthcare.gov and released the same PoS that was actually released? It'd be even worse than the current fallout because they aren't subject matter experts. At least OR appeared to try and be competent by having an in-depth sales cycle and outside consultants to vet the proposals.

ddingus 12y ago
Yeah, that's due to a strong bias away from anything else. Lots of lobby dollars in play reinforcing that, and there is the strong perception that Oregon can't actually afford to make any kind of investment. But this shows Oregon can, did, and it was botched horribly. Doing the same while providing local jobs, some great skills to be used all over the place, and build something that has a very likely lower overall ownership cost profile makes a ton of sense! It just won't get proposed in the current political climate, which was the point of my reply. Of course Oregon wants COTS. Everybody does, because they think the vendor will own it, saving them money through economies of scale. Quite frequently, COTS can do that, but for something like this, something new, that needs to communicate with a lot of varied systems? No way in hell it was ever gonna be COTS, and that right there should have triggered a much different, "how does Oregon do this?" kind of discussion. Oracle has a good sales team, and they sold it right to this dynamic, which is what they were supposed to do. Problem is the value really wasn't there. No matter what, it was going to take some real work to build out the exchange and have it function well. As for the support contracts, those come with an annual cost. It's very highly likely that cost could simply provide a job for the people who built the thing. Additionally, having seen success on a project like that, the same team would be well positioned to build out more and or get funded by other municipalities seeking similar solutions. I don't have the basic numbers a few of us ran back then, but they were compelling. This works the same for States as it does companies or anyone else. Make the in-house investments and own the tools. From there, buying help is always an option. Avoiding it is another option. Relying on a vendor, who will very frequently oversell their support staff, counting on their product to be a winner, often results in that support contract merely funding the trouble shooting. More dollars are required to buy, implement, test, etc... fixes. Fixes, which by the way, will be called custom patches for major customers due to the size and non COTS nature.
- ddingus 12y ago
  It could go badly either way. However, I find it very hard to believe a competent group would actually fail in this spectacular of a way. Vendors often know just a little more than the client does, and they do oversell, and "cost effective" gets calculated to a point where the purchase price and services costs are much closer to building something than not. And some of my career is in selling just that kind of enterprise solution, and that's exactly how it goes a lot of the time. There is value there for sure, and the vendor will need to capture as much of it as it can in revenue, while delivering as little service as possible to keep margin. Those expensive people, who likely would have built the system, wouldn't be doing nothing. Not at all. Once that system is up and successful, they continue to build lots of things OSS style. Boring, simple, cost effective things and anybody anywhere can pick them up and use them too. Lots of synergies here for States and municipalities looking to create jobs and manage costs. Having been on the ugly end of these things, it's worth a look. Perhaps it wasn't the right call for this project, but it really wasn't evaluated. And there are few entities able to answer an RFP to make that evaluation, who don't also sell very expensive software licenses. Should have been evaluated, and Oregon should have some basic competency in State IT to do that. The savings potential is significant, and Oregon needs to own it's IT anyway. For a lot of things, vendor supplied software will win out. No question. But for those things where it won't, or cost of ownership over long periods of time do not make sense for the public, having a means to explore that was what the legislation was all about. Keeps vendors a bit more honest too.
ddingus 12y ago
I agree with you on "tried", but this isn't the first State failure with that method, nor the first one with a lot of digits attached to it, though this one is the biggest I know of. Last time I did this, I hired a subject matter expert, or was that expert. Then the interests are clear, and it's not just a paid service. Oregon really should bring that competency into the State proper so that this kind of thing gets owned, not just sourced and managed. Do that with whatever software is best indicated too, but own it. Oregon didn't do that, again...
perlgeek 12y ago
> I'm sure they could do this project with OSS products and custom code integrating everything, but those don't come with support contracts. I'm sure that if you pay enough money (and considerably less than what they threw at Oracle), you can get support contracts for all kinds of open source software and custom integration solutions. That's the beauty of open source: you don't have to be the vendor to support it.
Confusion 12y ago
The main problem is that there is no such thing as 'a COTS solution for a water billing system with little custom code'. The only thing that exists is what companies like Oracle and SAP market as such, claiming their solution only requires 'configuration'. They confuse their customers with misuse of terminology, because in fact that 'configuration' consists of a lot of custom code that is needed to adapt their solution to a specific purpose. Whether it is more or less code than honestly and transparently assembling such a system from available OSS components is besides the point as long as customers are not crystal clear on the fact that COTS solutions rarely exist, except as components for a purpose-built system. A claim of having a COTS solution is almost always a sham for any reasonably sized project.

by joshAg 12y ago

ddingus 12y ago

Yeah, that's due to a strong bias away from anything else. Lots of lobby dollars in play reinforcing that, and there is the strong perception that Oregon can't actually afford to make any kind of investment. But this shows Oregon can, did, and it was botched horribly. Doing the same while providing local jobs, some great skills to be used all over the place, and build something that has a very likely lower overall ownership cost profile makes a ton of sense! It just won't get proposed in the current political climate, which was the point of my reply. Of course Oregon wants COTS. Everybody does, because they think the vendor will own it, saving them money through economies of scale. Quite frequently, COTS can do that, but for something like this, something new, that needs to communicate with a lot of varied systems? No way in hell it was ever gonna be COTS, and that right there should have triggered a much different, "how does Oregon do this?" kind of discussion. Oracle has a good sales team, and they sold it right to this dynamic, which is what they were supposed to do. Problem is the value really wasn't there. No matter what, it was going to take some real work to build out the exchange and have it function well. As for the support contracts, those come with an annual cost. It's very highly likely that cost could simply provide a job for the people who built the thing. Additionally, having seen success on a project like that, the same team would be well positioned to build out more and or get funded by other municipalities seeking similar solutions. I don't have the basic numbers a few of us ran back then, but they were compelling. This works the same for States as it does companies or anyone else. Make the in-house investments and own the tools. From there, buying help is always an option. Avoiding it is another option. Relying on a vendor, who will very frequently oversell their support staff, counting on their product to be a winner, often results in that support contract merely funding the trouble shooting. More dollars are required to buy, implement, test, etc... fixes. Fixes, which by the way, will be called custom patches for major customers due to the size and non COTS nature.

ddingus 12y ago
It could go badly either way. However, I find it very hard to believe a competent group would actually fail in this spectacular of a way. Vendors often know just a little more than the client does, and they do oversell, and "cost effective" gets calculated to a point where the purchase price and services costs are much closer to building something than not. And some of my career is in selling just that kind of enterprise solution, and that's exactly how it goes a lot of the time. There is value there for sure, and the vendor will need to capture as much of it as it can in revenue, while delivering as little service as possible to keep margin. Those expensive people, who likely would have built the system, wouldn't be doing nothing. Not at all. Once that system is up and successful, they continue to build lots of things OSS style. Boring, simple, cost effective things and anybody anywhere can pick them up and use them too. Lots of synergies here for States and municipalities looking to create jobs and manage costs. Having been on the ugly end of these things, it's worth a look. Perhaps it wasn't the right call for this project, but it really wasn't evaluated. And there are few entities able to answer an RFP to make that evaluation, who don't also sell very expensive software licenses. Should have been evaluated, and Oregon should have some basic competency in State IT to do that. The savings potential is significant, and Oregon needs to own it's IT anyway. For a lot of things, vendor supplied software will win out. No question. But for those things where it won't, or cost of ownership over long periods of time do not make sense for the public, having a means to explore that was what the legislation was all about. Keeps vendors a bit more honest too.

ddingus 12y ago

I agree with you on "tried", but this isn't the first State failure with that method, nor the first one with a lot of digits attached to it, though this one is the biggest I know of. Last time I did this, I hired a subject matter expert, or was that expert. Then the interests are clear, and it's not just a paid service. Oregon really should bring that competency into the State proper so that this kind of thing gets owned, not just sourced and managed. Do that with whatever software is best indicated too, but own it. Oregon didn't do that, again...

perlgeek 12y ago

> I'm sure they could do this project with OSS products and custom code integrating everything, but those don't come with support contracts. I'm sure that if you pay enough money (and considerably less than what they threw at Oracle), you can get support contracts for all kinds of open source software and custom integration solutions. That's the beauty of open source: you don't have to be the vendor to support it.

Confusion 12y ago

The main problem is that there is no such thing as 'a COTS solution for a water billing system with little custom code'. The only thing that exists is what companies like Oracle and SAP market as such, claiming their solution only requires 'configuration'. They confuse their customers with misuse of terminology, because in fact that 'configuration' consists of a lot of custom code that is needed to adapt their solution to a specific purpose. Whether it is more or less code than honestly and transparently assembling such a system from available OSS components is besides the point as long as customers are not crystal clear on the fact that COTS solutions rarely exist, except as components for a purpose-built system. A claim of having a COTS solution is almost always a sham for any reasonably sized project.