Issue link: https://iconnect007.uberflip.com/i/553274
August 2015 • The PCB Design Magazine 33 Viklund: It has to be an intelligent process where your data maintains design intent and intelli- gence of the circuit itself. You cannot translate it into dumb graphic and then go back again, because then you've lost all the intent. So that has to be maintained. Matties: that's what you were just talking about, maintaining the access into the data. Viklund: Yes, it's actually access into the data and intelligence of the data at all points. It has to be done that way. That's very important, and the other thing is that we cannot just dump the file, give it to somebody and tell them to im- port it. Because what we would be seeing is that every time you have file-based interfaces, one time or another a user will import the wrong version of the file and there's no way to see on the file that it's the wrong version, unless some- body realizes that the part is missing or some- thing. You typically discover it once you build the board, and that's really extensive. Matties: Dammit. Viklund: Exactly, dammit. We're done with that, we don't use files anymore. We open the tools and we have them talk on an Internet channel between them. The tools are basically commu- nicating back and forth live without files. Then they can use all the data management tools you have on either side to make sure that what you have opened is the right version. Matties: Would this be considered a cloud-based situation? Viklund: No, it's not really cloud-based. It works over the Internet but it's only the two tools in- volved in the communication, and it's kind of important because a lot of the times the users work with classified equipment and cannot use clouds. The cloud would have to be built inside their own organization like a mini cloud. But we don't see any advantage in this because it is two sides of the same design that need to talk for a limited period of time and then they don't need that communication anymore. So the cloud really is not that beneficial. Matties: so it's residing on multiple computers at whatever step in the process it's living in, but it goes back and forth immediately? Viklund: It goes back and forth. Typically you have a data management solution that makes sure you have the right version of the assigned database. That manages the work-in-progress data. Matties: When a customer comes to you, what's the greatest challenge that they're placing on you? Viklund: The greatest challenge is to really cut down the design time and get rid of iterations. Matties: are they looking for more autorouting tools? Viklund: Not for RF and microwave. For high- speed digital, definitely, because designs are getting so complex that even if you can route them by hand it takes too long. We have a new router that we call the sketch router, which basically allows the designer to sketch his in- tention on the board and the router will allow it to magically follow those guidelines so that the end-result looks hand-routed. That's very powerful. Matties: in my mind, it's like when the calculator came out. We really didn't need to know math as much as we needed to learn how to input into the machine to get the result we wanted. When i look 10 years down the road, is that what circuit design is going to look like? Calculator operators. not to take away from the experience and expertise of a designer in any way, but it seems to me that we put in parameters and there's a lot of intelligence compiled in the database. Viklund: It's definitely similar, because you have the problem now with having enough design- ers who have broad enough skill sets to manage these designs, so the tools will help the design- ers not get spread too thin. But I think we're still going to need intelligent and well educat- ed designers. We have to outman the designer more than replacing him. feature MENTOR GRAPHICS HELPS BRIDGE GAP BETWEEN PCB AND RF continues