Show HN: CADWOLF – Web-based math and engineering platform | Hacker News Reader

Show HN: CADWOLF – Web-based math and engineering platform

cadwolf.com

26 pointstheuttick19 comments

Submitter here,

I am a long time everyday reader of HN but I have never posted or really commented. I am a structural engineer with a few years of work in the space program. I have spent a great deal of time building something that I think can redefine the way we do engineering and I am close to finished with an initial version.

I am extremely nervous about posting this, but I would love to hear what everyone thinks. (There is a video in the overview section)

Thanks

Interesting. I would note the following:

"Applications are open for YC Winter 2016"

I have applied several times, including this round. I am a single founder and I understand hesitance to take people in that situation. I am trying to recruit some co-founders but it has been difficult finding the right people.

Hmmm. I might be interested in getting involved. Need to think about it a bit. I already run a couple of businesses and wasn't looking to add to that workload.

I happen to have an aerospace project coming up that will be quite intensive. I won't use your tool for the project but I will think things through as the project progresses in order to gain a better understanding of what your perspective might be.

My first reaction was that I liked the intent. Anything can save time and streamline a process has value. I can't comment on the merits of your approach at this point.

I see your email address is on the site. I'll contact you towards the end of the year if this makes sense for me.

Sounds good. If you're doing space stuff in houston, I likely have drank with someone you're working around. Be interested to know what you're working on.

I'll shoot over an email.

I used to work in a company involved in building CAE software and solvers. I can see value in what you seem to be doing but I couldn't figure out what. I should apologize that I didn't spend enough time but a short description would be nice. It seems like you are targeting very broad set of problems. It's usually hard to communicate that way.

I know it is difficult to describe complicated software simply. Make the examples more prominent, and provide a sandbox for someone to try them. Some people like watching videos, others don't.

There actually is an example document that any user can fully edit. It's under the docs and examples link.

As a mech engineer doing mostly structural work these days, you are definitely solving an important problem, one which I both witness and experience first hand every single day. Excel is used a lot, but it's error prone and sucks for communicating your work (formulas and logic is hidden). Plus, trying to track down the latest version of some calculation someone did 10 years ago, digging down 29 subfolders on the common drive, ugh.

I do like projects like Jupyter (I used IPNB a lot during my masters), but it seems like 95% of my fellow engineers feel like python/julia/even matlab are "too complicated for something excel does". Which I disagree with (ever tried extracting eigenvalues in excel? or simply plotting a hisogram?), but eh. Programming is also less visual/graphic, which is one plus of excel (colors, fonts, layout of a spreadsheet do help with processing information)

I can't say if your solution is "the one". It's too complex a problem for that. But my personal feeling is you're tackling the right issues at the right angle. As good as your product is though, getting traction is going to be the hard part, by far. I wish you the best, and certainly hope I get to use your product some day.

I agree. There is a huge amount of "inertia" in the engineering field. People find something they like and then use it forever.

If this has trouble getting traction in engineering as originally designed... Please pivot!

I watched the document section of your introduction video and was struck by how elegantly it would solve my needs for a blogging platform. I may even try using it for that without a pivot.

Interesting that you say that. There is a spot on YC applications that asked if you had any other ideals that you may have applied with or something along those lines. I mentioned pivoting to just that sort of thing.

I think this is very interesting, although I'm still trying to understand it. So excuse my ignorance, but how does this differ from open source projects such as Jupyter, Sage, and JuliaBox?

Or, as a commercial product, from the functionality available inside Matlab, Maple, or Mathematica? Those all have the same kind of notebook feature, but have the additional advantage of thousands of third-party libraries.

edit to add:

This project seems to be pretty neat, and it’s clear that a ton of work has gone into it, but I’m unlikely to ever use it.

Here’s my list of criteria for something like this:

(1) Is it open source? Am I being locked into someone’s proprietary environment? Can I inspect the code of every function I’m calling on, to make sure I trust them to behave correctly?

(2) Does it interoperate easily with my existing data and workflows? Can I easily scrape a webpage to feed data in, or import from a file in some obscure format from the 90s?

(3) Can I run it offline? Can I share it with anyone I want, and can they build on what I’ve made without paying a license fee? [Personally I think running on a computer locally is a big advantage over having everything done “in the cloud” and displaying in a web page, and thus being completely unavailable when offline. I’m assuming all the numeric code is being run on some server somewhere? Or is all the processing done locally in Javascript?]

(4) Is there some well tested library that already solves my problem that I can run within the tool, so that I can just download the library, use its API, and be done without a bunch of custom work on my end?

[Or more generally: does the tool solve some problem that I can’t easily solve in any other way?]

(5) Is the plotting functionality easy enough to make plots of data with a few seconds of work, but flexible enough that I can make a custom data visualization if I need to? (For a web project, I would just feed the data to D3.)

Don’t let me discourage you though. It’s great that people are working on tools, and quite possible that you’ll find a good market of folks who would use this thing.

Matlab, Maple, etc are very good at what they do, but they have two problems that my platform solves.

1. Your manager and/or customer can't see your work unless they have a license and can understand the language. Mine only needs a web browser.

2. If the analysis changes, the documentation must be updated as well. Mine does everything on the same web page.

As for the questions, those are all ones that I have asked myself about several times and here is where I am now.

(1) It is not open source and you really are kinda locked into using the system once you get going. I really couldn't find a way around this as the real appeal for most engineers is the absence of the need to code stuff. I have documentation that describes how I solve most algorithms and some demos.

(2) Scraping web pages is high on my to do list. You can paste most forms of data already, but you can't yet import files directly. Seeing that this is high on someone else's list I may put most of my effort there.

(3) Right now, everything is done locally in javascript but there is no way to save changes without server access. Part of my business plan is to use a server side code to solve large data sets and possibly charge for that. I have been surprised at how well the browser has handled most problems though.

(4) The main problem that this will solve for most people is this ... You get data from a test run somewhere. You put that data on a server, run a Matlab or Python code on it and then create a word document to pass on that info. With my system, you upload the info, analyze it, and display the results all on the same web page. It isn't just about analyzing the data, it's about doing everything on the same platform.

(5) The plotting functions are pretty easy to use and you can get something up in a few seconds. I also have 3D plotting.

My main suggestion would be to first find some concrete customers, and then ask what they need. (As I said, I’m not likely to be one of them, so don’t prioritize just based on my opinions.)

All the best luck!

Thanks for the input and I totally get what you are saying. It may help to know that the end goal is to tie in a CAD system. This way, each "document" contains the math to design a component while the CAD reads from that to display it. This way, the CAD model automatically updates with changing design requirements just like the part trees already do. What you see is really just a first step.

I like those projects. What I am doing differs in purpose and function

Those platforms require some sort of installation and are usually meant to be tools for a programmer to create a web page that utilizes existing code like python, etc. My platform has no coding in that sense. You create a new file by clicking an icon and add blocks of text, equations, etc through a graphical interface. All the user would enter from a code standpoint would be something line "a=sin(10x)+cos(3x)".

Whereas Jupyter and others are tools for programmers to use, my platform is a tool for other people to solve and document problems. They target programmers whereas I target engineers and scientists. With them you are writing code and with me you are using a word document that also solves equations and makes charts.

Reply on news.ycombinator.com