Response to SciChart WPF Survey July 2016

So a week ago we announced our H2 2016 Roadmap for SciChart WPF and also a survey on what features you would like next. The response has been really positive! Lots of people giving feedback and giving us some good direction for the next half of the year.

The feedback was totally anonymous so we can’t actually reply to any of your comments, so we’re posting some responses below. If you want to get in contact with us to discuss any of these further, please do! The more dialogue we have between you, the customer, and ourselves, the better.

Q: What Chart Types to build next?

Almost everyone wants a Pie Chart. Really? OK then! These are actually really simple to build, so much so you could actually do it yourself, but we’ve avoided it until now so we can focus on the specialist area of high performance scientific charts. However, since its a quick job we’ll try to fit it into the next iteration.

We also have some good feedback about what specialist scientific charts you would like to build, so thank you for that.

Q: Which platforms do you want?

We have some great feedback on what platforms you would like us to port to. However, with so many platforms to choose from, and the cost of one platform port being high (you are looking at 3-4 developers for a year for one platform), we have to be careful which to choose.

For instance, we have recently ported to iOS/Android so Xamarin makes perfect sense. In fact, Xamarin is a case of simply adding a binding wrapper around our existing iOS/Android controls so the two go hand in hand. It’s likely we will look at this soon.

However, porting to HTML5 is a complete re-write. HTML5 we can see a big (yet crowded) market, however you’re aware that we’re not going to achieve the same performance as in WPF right? Javascript by its very nature loses 90% of the speed of the CPU and there are no nice workarounds to use C++ or to write your own drawing algorithms!

Q: Platform port Sponsorship

If you were one of the people that said you wanted to sponsor a platform port, please get in contact. We have the ability to scale our team quite rapidly and if we have some certainty of income it means we can hire a dedicated team and get to work.

Q: Optimised Algorithms for Descending Charts

Two people asked in the comments for us to optimise the algorithms for descending charts. As you know, SciChart excels in performance when the X-data is sorted ascending as this allows us to use a host of optimised algorithms for indexing, resampling and drawing. But what about sorted descending? SciChart treats this as ‘unsorted’ and drops down to a slower drawing path.

However, did you know that there is an optimisation in SciChart already to handle this case?

Simply append your data to the DataSeries sorted ascending in X, and use XAxis.FlipCoordinates = true. This will render your data as descending while leveraging the full power of SciChart’s accelerated drawing.

If you have a comment about this, please use the comment box below or contact us.

Q: Exporting to a vector based graphics format

One person requested SciChart WPF to export to vector for seamless inclusion into HTML5 or PDF documents. However, are you aware that SciChart already supports export to vector graphics? You can see a demo in our Export to File Options example.

SciChart supports export to PNG, Bitmap, Jpeg and now XPS (vector) for high-quality, scalable output. The XPS output is available in the SciChart WPF 2D Enterprise and SciChart WPF SDK Editions only.

Q: DateTime axis synchronisation within SciChartGroup. At the moment we need to add Double.NaN points in order to achieve this.

We are working on a new Axis type called a DiscontinuousDateTimeAxis. This is a value axis like DateTimeAxis which allows you to exclude certain times, e.g. after market hours, weekends. However, its behaviour is like a DateTimeAxis so it obeys Date value for positioning of points.

This should solve the above problem, however, the performance of DiscontinuousDateTimeAxis will not be as good as CategoryDateTimeAxis. This is because of the higher number of calculations required to allow this type of axis transformation.

We hope to have this axis type ready later this year.

Q: Better possibility to extend existing charts with custom series.

..Example: in a ohlc or line charts we want to visualize our market activity, eg plot our orders as horizontal lines

We do actually have a CustomRenderableSeries type which allows you to draw virtually anything. Can this help? There are some good tutorials in our documentation plus an example in our Examples Suite.

Q: I think some bugs are long overdue.

If you want these fixed, you should really tell us what they are 😉

Maybe you have in the past, but we get so many requests for enhancements or fixes that its worth telling us again to remind us of the importance to you.

Q: I think SciChart could benefit from having an out-of-the-box histogram

… instead of having to do work-arounds to make it work. I could provide examples of things we’re doing that would add value to this new histogram series.

Yes please do! Our business-knowledge of statistics will not be as good as yours. We are the chart drawers, you are the chart users. If you want to supply us with any examples or code that we can build upon in any business area, please do!

Q: Proper MVVM Tutorial

One person requested a full MVVM tutorial. It’s difficult for us as we have to show people how to setup a chart in code (fully code), or in MVVM, or in a hybridised sort of MVVM depending on what their level of experience/comfort is with the MVVM pattern.

But I guess we could do a full reference implementation of dynamic everything (series, points, axis, modifiers, styling) with MVVM. It’s a lot of work for us though, hence why we haven’t done it until now.

Q: Include Examples in Source Code Repository

Good idea! We can do this quite simply. I will log a task on the board!