Is Eagle retired?
Hi Everyone & for those in the US, Happy (belated) Thanksgiving! --
The decision to focus on Fusion over the last 18 months is something, as a team, we stand-behind but I want to stress, there are two discussions here. Feature development on EAGLE vs. Fusion and the discussion of ‘Cloud’ — which for some of you are intertwined but which, as Jorge mentioned, is something where the narrative is changing as more and more users find customers who understand that vital data is invariably stored online and this is a “trend” that’s indeed become something of a new-normal. That said, I dont want to try and convince anyone what their customers require because that isn’t fair to any of you, so I will focus on the EAGLE vs. Fusion discussion.
Our decision to focus on the Fusion capabilities right now are not based on the move to cloud-storage nor is “cloud” the reason feature development has been slowed or stopped on EAGLE in recent months while our efforts focused on Fusion Electronics. Instead, this is based entirely around limitations in the core EAGLE code base (what we call Table Stakes) and limitations in its rendering, selection methodology, command-framework, UI, data model, etc. These are the reason we are instead focused on Fusion today. Moreover, we will continue to develop on the Fusion side until such a time that we prove-out (for users) why Fusion’s core is SO much stronger in multiple areas and by implementing the things which demonstrate “why” we’ve made this choice with the resource constraints any team has on any product. We aren’t there yet, but to say “EAGLE is dead” implies you can’t continue to use EAGLE today and though it is not getting new feature development at the moment, we made the decision to keep EAGLE going until we knew clearly that what we believed were best in-class capabilities (which we need for any future development) would work for PCB and the users were able to make the switch.
These additional capabilities will enable us to make substantive changes that drive real transformative capabilities we know from conversation with the community and our team's long time in this industry, are valuable to every one of our users. To achieve this, we need to balance time focused on future direction- and time spent on current capabilities built on this new tech stack -- leaving little time for back-porting capabilities from Fusion to the simpler, less powerful EAGLE core which can't support them without creating something nearly identical to Fusion anyway. Moreover, trying to do both only prolongs what we are trying to do in Fusion to demonstrate what we shared a year ago about performance, selection, snapping, rendering, routing, dimensioning, etc and we can’t delay this or we dont have a strong position to establish new, ground-breaking workflows which make the investment in EAGLE relevant for the *next* 20+ years. Jorge’s comment on cloud adoption only signals that for many, if not most designs, we see users utilizing cloud capabilities in various forms, but these aren’t the primary driver of why we are doing this.
To be sure, Jorge, Edwin, Richard, myself and others are deeply familiar with EAGLE and the engineering team today are familiar with the EAGLE code base. Development on EAGLE is "easier" for everyone on the team (today) - owing to a more basic, raster-only, bitmap rendered general-purpose CAD core. Moreover, it would make our lives far easier if we didnt face competition from other tools with core capabilities developed more recently, by a much larger engineering team than what EAGLE had pre-Autodesk and able to make better use of today’s compute resources.
Simply put, the EAGLE graphics & geometry engines / data model / memory management are not suited to high-performance, high-complexity designs and yet, the 2D graphics capabilities Autodesk owns and has proven year-in, year-out with Autocad and other tools like Fusion 360 and Inventor are-. Those capabilities are truly second to none and are built on 40 years of geometry know-how that almost no other company on the planet has like Autodesk. I’ve personally worked at several companies creating ECAD tools for ‘the future’ and as yet, I have seen no other company better positioned to use what IP they already own to fundamentally & positively impact design & manufacturing. So over the last 12-18 months we have been steadily at work focused on a strategy split between adding capabilities on top of the foundations we laid in Fusion and strengthening the core with the best of Autodesk to where we can easily handle a much larger PCB, with greater performance, usability and scalability. It is neither easy or even likely — to simply “port” those features to EAGLE and if you had a view into the core of EAGLE you would understand why what EAGLE has is good but not going to scale.
Have a quick look at the images below paying particular attention to Autodesk's geometry library, selection framework and command framework...
In the images above, what’s notable about every one of these is that they depend on a complex combination of linear algebra, vectors, constraints, parametric input, and of course performance optimization. This isn’t easy to implement, inexpensive to build (it has 40 years of Autodesk investment behind it), yet it is truly game-changing for the challenges people face in EAGLE, Altium, Cadence, or Mentor today. Sure, a background in Technical Drawing and a pencil might get you the line endpoints you need to make complex geometries, however nowhere in the current EE curriculum, do we focus on Technical Drawing (I still enjoy teaching periodically) and moreover, the EE landscape demands faster iteration, greater accuracy and with that, better workflows to enable the marriage of Industrial Design, Mechanical Design, Electronics Design, SW Development & Manufacturing. Architecture and Construction have problems at a scale every bit as complex as PCB (think: Burj Khalifa, Bay Bridge, reconstruction of Notre Dame) and Autodesk has the technology stack which makes these things possible if we integrate the right mix of these core services and reimagine them applied to PCB.
Geometry is only one slice of this. More interesting is when geometry isn’t the input vector, but rather Domain-representation of the data like Tpd (Length), Capacitive Coupling (Spacing, Frequency, Cross Sectional Area), Current (Cross Section). Converting that to a geometric result (CAD) in the interval you expect the mouse to respond is where mathematical performance is King.
At this stage we have the team focused on the Fusion implementation because in betting on the future, we will bet on Geometry, Constraints, a better, more capable Rules system, better rendering, selection, graphics performance, hit testing…and as we truly believe, a better overall customer feature<>capability<>experience mix. To put any of that back into EAGLE is a) altogether not possible and b) not what is in the best interest of the users long-term, if we want the core of software to truly help you address design challenges in a way that makes both you and Autodesk competitive in Electronics. If we fail at that, then we needn’t worry about EAGLE’s future because that fate would be sealed. So we choose to press forward with this and we will succeed in this for our users and our team and our families, as well as the hope of making EAGLE as it is today, a commercial success against the big 3 or 4 players on the market. Your support will fuel that and though some of you may not support it, for those that do- just please know that it motivates us to work faster and work harder when we hear that you see something you wish we had in EAGLE which we’d added in Fusion but you’re not ready to move or you just completed a design in EAGLE + Fusion that demonstrates what’s possible today that wasn’t in EAGLE v7.
Best regards,
Matt
Director of Engineering
Fusion Electronics, EAGLE, Tinkercad
Hi Guys,
Firstly thank you all for your detailed responses. I read them all and I hear what you’re saying 100%. No question, you made great points on the cloud side of things and it wasn’t to dismiss the discussion around 'cloud versus desktop' but instead to focus on the things which are available in Fusion & why they help us build the tools we want for users. It doesnt invalidate the cloud argument(s). However, after 5+ years in the EAGLE code and the myriad things we had to do just to get Table Stakes like solid polygons, push and shove, obstacle avoidance, observer-based DRC, Inspector, Design Manager, Quick Route, ECAD/MCAD, etc; we know where the bones are buried and we know how much further we could take EAGLE. The dilemma we face is that to achieve commercial viability and build a competitive tool (more powerful, performant & innovative) and the decision was made to use the technologies within Fusion because it is what we need to deliver the next 10 years of PCB software. Moreover, if we dont create a tool with technology required to deliver features comparable to the big-3 or big-4 competitors, we only hasten the outcome nobody (us included) wants.
To be sure, we have heard that the cloud component is what makes a number of folks unable or unwilling to move to Fusion & many of you would prefer local file storage. However we also see a lot of unrealized value in connectivity to the back-end that enables more distributed workflows / distributed compute resources. Today that isn’t obvious but I see a LOT of opportunity for things we can do with BOM, Mechanical Drawings of PCB (including the sign-off process, redlining, where-used, etc), CAD/CAM/CAE & Mfg interfaces, route, sim, etc. but I wont suggest we are there yet. Still, there is a price we pay up front for what’s a long term vision that can look confusing on the surface but there is a method to our madness, we just aren't there yet. Like @RitchieFromParis and @Anonymous-MESO say very accurately, many things are still far too confusing where the cloud was introduced. Honestly, that’s a bunch of long-time C++ developers with deep knowledge of low level coding cutting their teeth on what is possible and navigating our way around a very mature, very robust back-end. What is truly helpful about this feedback is it gives us something to focus on first, ie. “if you do ‘X’, then maybe, just maybe I would consider it”. For some of you, we may never convince you or you may never be able to- but we'd no less like to capture the definitive list of reasons "why?" and where your customers are giving you the directive, what their concerns are so we can attempt to address them.
On a lighter note, a bit of a teaser about a common issue with Cloudy-stuff that we have in the works — the major one being library-related. We knew that libraries and cloud “content” are more confusing when you have to ask the question “where am I putting this and why do I need to know where I’m putting it? Dropbox doesnt have 3 different Dropboxes, why do I have 3 different storage paradigms?” Hearing that, we set to work and we are close to wrapping up a library consumption experience (ie a replacement for “Add Part”) that is non-modal and which provides a single view across all assets you own, whether locally, on Library.io or in Fusion Team. This will cut down on the duplication and confusion and make sourcing and placing parts within your part libraries much easier. I’m hoping we will have this out in the first part of next year but everything is “TBD” until it’s released so dont think we’re waiting on anything but stability & performance.
Speaking of performance, a number of you picked up on the slower behavior you saw in Fusion. This is our other, highest priority and like October, the November release / early December has some stuff that continues to address things like ULP performance, “observer” performance (observers fire anytime something changes which handles notifications of Panels, Lists, Rules, etc.), polygon performance / results (esp for manufacturing outputs) and a lot of little things we found in very old code that made everything slower (read: a for-loop that has a single item in the index…always? Some of this was very old code that was written before the days of profilers). Pay close attention to the What’s New? document and subsequent blog post for the next release which highlight some of the work we’ve been doing to make performance significantly better (our tests with ULPs from Oct and later showed ULP performance at least as fast as EAGLE 7.x and even faster in many cases).
Finally, a quick note on cloud capability. Not sure how many folks are aware but it’s worth sharing a quick gif that shows how data can be shared and navigated in the online viewer we call Fusion Team. Reason I share is that I hand solder a lot of prototypes and I get into a state at the bench where just finding parts, testing signals, etc means I either need Fusion open or thru Fusion Team (included with Fusion) I can just have a lightweight laptop on the bench or even tablet. Some other lesser known things are the Where Used in Fusion Team and even the addition of Properties, exploded views, etc. Still not perfectly implemented IMO for the EE persona but there is a lot of useful goodness there which isn’t obvious because most users never know they have Fusion Team when just using Fusion (this is on us to make this more prominent).
Thanks everyone for the feedback. I don't get to make ALL of the decisions but I want to stress that we always want to make the right decisions with the things we have influence over and in the interest of growing the community of users thru increased capability. In some cases, we make long term choices because that’s what new feel is best (y)our future success. Within the scope of what I can influence and what we control; I will always press for feedback because we have to deliver something which makes this-part of Autodesk’s portfolio a commercial success (which in turn drives resourcing, ergo more development, ergo more capability and greater influence on product planning and roadmap). If we fail at the commercial side of this, then indeed, EAGLE is dead. (Suffice it to say, Fusion Electronics is growing — but we want to avoid anything which proves a roadblock if we can't deliver what we need to be competitive.)
Hope that helps and hope I didnt leave anyone feeling like I’m missing the elephant in the room. I’m not one to shy away from conflict where I can influence something and where the product team has chosen to prioritize something.
Best regards
Matt - Director
Fusion Electronics, EAGLE, Tinkercad
Hi @engineering Hi @jeffg28CLY ,
If you search the Fusion forums you'll see that the ITAR and the cloud situation is a general concern to Fusion 360 and it's something that's frequently inquired about. As Matt mentioned we hear the concern and the Electronics team is a voice within Autodesk pushing for some resolution to this issue. Whatever gets implemented on this front will affect the whole of Fusion 360, not just Fusion 360 Electronics so it will require buy-in well beyond just the Fusion Electronics team. All this to say, the Fusion 360 team is aware of these concerns in general and I'm confident that at some point something will have to be done to address this. Whenever it's done it will be an effort guided by the overall Fusion 360 leadership not just Electronics.
In regards to the forced updates, that's an internal discussion that's been going on for a while with many different solutions proposed. Again, like Matt mentioned in what the Electronics team can decide we'll push for as much flexibility for users as possible. With that said, any change on this front will also be guided by the general Fusion 360 leadership since it affects the whole of Fusion 360 and not just the Fusion 360 Electronics team.
I encourage you to post the cloud concerns and the force update concerns to the Fusion 360 Design, Validate and Document forum as well, since it will be seen by a larger Fusion 360 community (MCAD engineers and modelers who share similar concerns) that can also pile on their voices to that thread.
Thanks for your feedback, we take it very seriously and disseminate through the larger Fusion 360 org as much as we can.
Best Regards,
Jorge Garcia
Product Support Specialist for Fusion 360 and EAGLE
Kudos are much appreciated if the information I have shared is helpful to you and/or others.
Did this resolve your issue? Please accept it "As a Solution" so others may benefit from it.