There is now too much data system innovation - there are scores of projects each of which implements a particular idea and very few of which have a large enough community to move beyond version 0.3
We can't put our pipelines onto these because the cost of understanding and adopting them overwhelms the benefit of ditching our internal code. We are in the same situation as the NPM / Javascript folks, but our data is an enterprise asset and we have to consider the impact of a project just going away.
The vendors have held us to ransom for years and we've had to break out because it had got to an industry destroying level - really the bills are material to the stock price and time and again it turns out that despite handing over $100m the features and machines you need are not included and you need to send just another $5m over. The money stopped being used for R&D when the consolidation of the 00's happened. The vendors set themselves up as vertically integrated solution providers with the technology as a lock in factor rather than a competitive differentiator. This would have worked if nothing in the economy ever changed again but oddly it turns out that our needs are radically different when competing / collaborating with Facebook & Google vs "old corp who has gone bust now" who were our traditional enemy.
So we're in a fork. The old route of using a trusted technology partner has gone, they have all betrayed their customers every quarter for the last 20 years, not only can't we trust them but we can absolutely predict when and how they will screw us. We flag it for every project to our execs, it's a built in assumption. On the other hand the big hope of opensource alternatives is not arriving in the way that Linux arrived, or maybe it's arriving in the way that Linux on the desktop arrived.
The solution is that the middle tier of corporates have to get more real about opensource. At the moment there's a handful of companies in the Forbes 500 who are significantly involved. CIO's and CFO's don't see the benefit, there is no case at the moment, but if there were 500 >$1m opensource programs running in the corporate world rather than 20 things would be different.
There needs to be a standarization effort and a co-ordination layer. We also need to think through the value chain as well. At the moment leaving it to the market and "revenues from professional services" are not working so well, and I am very cautious about the trend to paid for bits and bobs on top of open source.
The worst outcome would be to go through all this and find that we are back to vendor hell but chained up in a different basement wondering where the bastards in suits have gone. Instead we're going to be guessing that the guy in a cool t-shirt and designer jeans who's fiddling with a blow-torch is going to want something soon.
Agree - it's a notional form of free that includes your CFO paying through the nose without realizing it.
The office up strategy is really good for MS. I've always wondered why they haven't made more of Exchange as a data platform given that they have a near corporate monopoly on that and there is vast latent value in the data that is in it.
There is now too much data system innovation - there are scores of projects each of which implements a particular idea and very few of which have a large enough community to move beyond version 0.3
We can't put our pipelines onto these because the cost of understanding and adopting them overwhelms the benefit of ditching our internal code. We are in the same situation as the NPM / Javascript folks, but our data is an enterprise asset and we have to consider the impact of a project just going away.
The vendors have held us to ransom for years and we've had to break out because it had got to an industry destroying level - really the bills are material to the stock price and time and again it turns out that despite handing over $100m the features and machines you need are not included and you need to send just another $5m over. The money stopped being used for R&D when the consolidation of the 00's happened. The vendors set themselves up as vertically integrated solution providers with the technology as a lock in factor rather than a competitive differentiator. This would have worked if nothing in the economy ever changed again but oddly it turns out that our needs are radically different when competing / collaborating with Facebook & Google vs "old corp who has gone bust now" who were our traditional enemy.
So we're in a fork. The old route of using a trusted technology partner has gone, they have all betrayed their customers every quarter for the last 20 years, not only can't we trust them but we can absolutely predict when and how they will screw us. We flag it for every project to our execs, it's a built in assumption. On the other hand the big hope of opensource alternatives is not arriving in the way that Linux arrived, or maybe it's arriving in the way that Linux on the desktop arrived.
The solution is that the middle tier of corporates have to get more real about opensource. At the moment there's a handful of companies in the Forbes 500 who are significantly involved. CIO's and CFO's don't see the benefit, there is no case at the moment, but if there were 500 >$1m opensource programs running in the corporate world rather than 20 things would be different.
There needs to be a standarization effort and a co-ordination layer. We also need to think through the value chain as well. At the moment leaving it to the market and "revenues from professional services" are not working so well, and I am very cautious about the trend to paid for bits and bobs on top of open source.
The worst outcome would be to go through all this and find that we are back to vendor hell but chained up in a different basement wondering where the bastards in suits have gone. Instead we're going to be guessing that the guy in a cool t-shirt and designer jeans who's fiddling with a blow-torch is going to want something soon.
It's up to us to work out a new model, but what?