HNHacker News
TopNewBestAskShowJobs

milen

4,570 karma · joined August 28, 2011

Software engineer on Build Infrastructure @ Facebook, ex-tech lead on Spark AR Studio, amongst other things created Monodraw, co-created Clear, EventBox and others.

More about me at http://milen.me/

submissionscomments
milen··on ASCII art editor designed for the Mac
Thanks :)

There's currently no colour support, as my focus has been to make a solid v1. I've looked into ANSI escape sequences to support colour and it should be quite easy to add.

Part of the problem when I work on Monodraw is that there are so many cool things I stumble upon related to text art and I want to do them all but have to restrain myself, as there's only so much I can do.

We do have something related to HTML / CSS output in the pipeline which I'll be talking about more over the next few weeks.

milen··on ASCII art editor designed for the Mac
I don't want to give any estimates as they would just be broken, lead to user disappointment and put unnecessary pressure on us.

But I can tell you that from the bits that we have left, none of them require any internal architectural changes and it's mostly just implementing missing bits and pieces.

We have two major improvements that are left to be implemented (I'm in the middle of the first one) and the rest are minor. For example, document zoom is trivial to implement (practically, it involves putting some UI controls and some small bits of code to tie loose ends).

I'll be very open about the development process and try to keep everyone informed as much as possible. My advice would be to follow @Monodraw, our blog and/or myself (@milend).

milen··on ASCII art editor designed for the Mac
No, we haven't settled on a price yet (I don't know what it should be). If you have any thoughts on the subject, I'm all ears :)

Our top priority currently is to create tool that we're proud of and which ours users love. The financial side is only secondary (although very important). When we get close to releasing v1, we'll put more thought into it. The beta will be free to use for everyone (and that should last quite a while).

The development of Monodraw will be very open from now on, I'll be writing about my thought process about all aspects, if you're interested make sure you follow our blog.

milen··on ASCII art editor designed for the Mac
It's quite easy for me to support arbitrary character sets, so if there's enough demand, I can definitely implement it. If you can just drop us an email (see "How do I provide feedback" section at the bottom), that'd be great - thanks.

But it won't be supported in the initial public beta build, as we have to prioritise other areas (sorry!).

milen··on ASCII art editor designed for the Mac
Apologies that the robot didn't render correctly in your browser - here's how it should look http://imgur.com/pst21UO

I believe this is caused by the font used to render it, which is not a monospaced. We'll investigate it first thing tomorrow morning and make sure it works on all platforms / browsers.

Sorry about that - thanks for letting me know :/

milen··on ASCII art editor designed for the Mac
I don't think all the cool tools are coming out for the Mac but it's undoubtedly skewed towards it. The reasons for this are a mix of cultural, social and technical, along with momentum.

Speaking for myself, when I was getting into serious programming, it was right in the heyday of the Delicious Generation (Delicious Library, Disco, AppZapper and more). Back in the 10.4 / 10.5 days, eye candy and ease-of-use were the top priority. Some notable apps came out and they in turn inspired other developers to follow the same path - a polished user experience is absolutely vital. This creates a self-perpetuating cycle, where the existence of such apps creates fans who dream of making apps like that (I'm one of these).

milen··on ASCII art editor designed for the Mac
Thanks! I replied to fensipens explaining that the app can work in both Unicode or ASCII mode, which can be switched on the fly.

The reason why I made the screencast in Unicode mode is simply because it looks nicer :)

milen··on ASCII art editor designed for the Mac
Thank you :)

Yes, we are going to charge for it because we have to - we are a very small independent software studio without any funding.

The future of the software belongs in the hands of our users, they will decide if our work is worth it and support us financially.

milen··on ASCII art editor designed for the Mac
I've designed the app so that you can choose your character set - Unicode or ASCII. This is a property of the document and can be changed on the fly, so you should be able to use the tool for text in a legacy environments as well.
milen··on ASCII art editor designed for the Mac
You're quite right that the most important thing about the tool is that you can take the exported plain text result and put it in your code, web pages, etc.

But when it comes to actually creating the content, we want to provide the best user experience. Monodraw can be thought of as a drawing app which was designed assuming the output would be ASCII.

In my view, those aspects are orthogonal - you can have a hard to use tool that exports into a standard format or an excellent tool that produces proprietary files.

milen··on ASCII art editor designed for the Mac
As far as the internal data model is concerned, there are no newlines. All shapes exist in an abstract discrete coordinate system and draw to it (if you're using the text tool, you can insert new lines etc but that's internal to the text shape, the canvas only cares about what characters appear at what abstract point).

Of course, when we export as text, we need to insert newlines so that it renders the way it's displayed.

Thanks for the kinds words about the execution :)

milen··on ASCII art editor designed for the Mac
Hey guys, I'm the developer of the app - glad to see it appear on HN. I'm working on it as we speak, so that we can ship the beta out ASAP.

If anyone has any questions, just shoot - happy to answers :)

Thanks!

milen··on The Financial Side of Building Mac Apps
Whichever route you take, you _must_ watch any relevant WWDC sessions - they contain a lot of valuable information, and more importantly, in many cases, information that's not available in the docs.
milen··on The Financial Side of Building Mac Apps
I don't know about your jurisdiction but in most EU countries, the sale proceeds from Apple lie outside the scope of VAT. Thus, you need to subtract your effective VAT rate and the net effective rate that you end up paying on the MAS is about 10-15%. Which is very close to what you would pay for, say, FastSpring ("You have the option to pay either 8.9% flat or 5.9% plus $.95 per transaction" [1]). The biggest advantage of the MAS is that it's a distribution channel preinstalled on every single Mac - that's extremely valuable.

As you say, you would ideally sell on both the MAS and outside. In terms of selling outside the MAS, it does come with some overhead, mostly related to VAT. Even worse, the EU is changing the VAT rules starting from 2015 (the rule change itself is designed to remove the competitive advantage of big companies to be based in EU states with low VAT and sell products in countries with high(er) rates):

> Currently, the place of taxation for digital services supplies is determined by your location as the supplier of the services. However, from 1 January 2015, the place of taxation will be determined by the location of the consumer.

It's not a big deal since you can just sign up for the VAT MOSS but it's still more hassle than just putting up an app on the MAS. Note that all of this applies after your exceed the VAT threshold, which is different in different EU member states.

[1] http://www.fastspring.com/pricing.php [2] http://www.hmrc.gov.uk/posmoss/

milen··on The Financial Side of Building Mac Apps
Even though your particular split between upfront + recurring costs is different than in my example, it's still in the spirit of sustainability. I honestly have no idea how the economics pan out if you had $50 app and $X yearly upgrade (or is it $X for 1yr of updates?) compared to, say, $20 each 1.5yrs. But in both cases, I think you'll have a chance of building something that can last.

Interesting stats about the lifetime upgrades - thanks for providing those. I guess it works as long as all of your users don't buy it (which means your recurring revenue drops to 0 at some point and you're back to square one).

milen··on The Financial Side of Building Mac Apps
I generally agree with you. One of the interesting bits of information from the article is the financial boost from paid updates. You can sort-of approximate this on the MAS but none of the options are ideal.

In my opinion, you absolutely must give people the chance to try your product. This means having a trial build on your website, which most Mac indies do but it's a bit problematic for paid apps, as I presume a lot of people on the MAS itself are unaware. It's in the best interest of everyone to have trials - for customers, so that they don't feel duped and actually buy a product they're happy with and for developers, so that they acquire real customers, not play a game of "can we convince people to fork over money without trying".

Moreover, it's in the best interest of customers to actually provide paid updates - they can vote with their wallet whether a business should survive. But without the proper availability of paid updates, customers are left with no choice, even if they did want to support developers. Unfortunately, the voice of a vocal minority instils fear in developers to try and charge for such updates.

Given the current climate, our plan for the future is to lower our initial price and offer major versions as paid updates. Instead of, say, charging $40 for the current version + $20 for an update, we'll just always charge $30 for the current version. As a side effect, it lowers the entry for everyone and your real customers can still vote with their wallets and support your efforts, if they choose to do so. One of the downsides is the introduction of a new SKU which is not that much of a problem (before the MAS, you had to do the same anyhow). It's usually wise to provide a way for customers to stay in touch with you and hear about updates but only if they want to (don't smack dialogs repeatedly in their face, that's just not nice).

milen··on The Financial Side of Building Mac Apps
I've been talking to iOS / OS X indies over the past few weeks and with very few exceptions, all I hear is "revenue down X% over the last Y years". We've seen this first hand, too.

Luc Vandal wrote a good article [1] about it. I'm very interested in hearing about other people's thoughts and experiences - contact details can be found on my HN profile.

[1] http://lucvandal.com/2014/07/03/theindielife/

milen··on Swift is not
I did some further investigation [1] by running 20 million iterations in Release and profiled the code.

Turns out, at least 82.9% of it was just swift_retain / swift_release (ref counting). Whether those retains / released can be optimised away, I don't know, we'll have to wait and see.

[1] https://twitter.com/milend/status/474360984413683713

milen··on Lookback.io - See how people really use your app
I seem to have missed the FAQ - thanks for the pointer. For me, a lot of value would be derived by having the ability for any user to send us a video of the problems / questions they're having (e.g., how do I do X, etc).

I realise that Lookback is mainly geared towards internal testing but it should be easy to aid the support workflow, too. Capturing the status bar is the only private API you actually need to use to fully recreate the screen, although I don't know whether the team behind Lookback is using additional priv APIs for other parts.

milen··on Lookback.io - See how people really use your app
Congrats on the launch, looks impressive and very nice.

What caught my attention was how you managed to snap the keyboard and status bar and get the app past approval (which is what I assumed from the demo). But after giving it a go, it seems that your SDK doesn't actually capture the kbd and status bar.

Being curious, I did some further digging into your SDK and UIKit. I presume for the demo, you must have used a custom SDK build so that you could tap into some private APIs? You can get the UIWindow for the keyboard via public APIs (UIApplication's -windows) but the status bar is private (_statusBar ivar on UIApplication; easily accessible with -valueForKey:@"_statusBar"]). Note that _statusBar is a subclass of UIView (direct subclass of _UIScrollToTopInitiatorView which is a subclass of UIView).

So for anyone else wondering how they capture the full screen, that's the way it seems to be done :)

Oh, by the way, if anyone's doing screen / view capturing, I highly recommend looking into the new iOS7 APIs (-drawViewHierarchyInRect:afterScreenUpdates: and -snapshotViewAfterScreenUpdates:).

milen··on App Store Pricing: Worth Less than a Cup of Coffee
I am the engineer behind both Clear for iPhone and Clear+. Your statement could not be further from the truth - it has taken a lot of effort to make the iPad component.

I'll give you a taste of some of the work that needed doing (completely describing everything would take ages). Here's a few examples:

- With iOS7, we switched to using TextKit so that we can vary the text sizes (both internally and via Dynamic Text). This involved a significant refactoring as in the previous codebase, we had to hardcode offsets. Why was it all hardcoded? Anyone who's had to deal with the problem of editable inline text would know that there were no APIs to match the drawing of text and editing via UITextView (which was previously based on UIWebKit; while you could the drawing either via the convenience NSString methods or via CoreText; if you used CT, you'd end up with mismatched visuals when editing / displaying). Did you also notice that when completing items on iOS6, the strikethroughs extended beyond the text on items with more than 1 line? Not anymore, due to the new internals.

- The architecture on the iPad is completely different vs the iPhone - on the iPhone, you have a single "list view" (custom implementation to be able to provide all the advanced gestures). On the iPad, you now have two view controllers that need to stay in sync, animate in sync, etc. It also meant that I had to do significant refactoring to abstract away all the controller logic as it had to be embedded in different controller hierarchies - you really don't want to duplicate all that logic.

- List peeking on the iPad. This required adding support to both the list view and to the item views, again - something many people might not even use but it's extremely handy when you need it.

- Drag and drop on the iPad. iOS does not provide a standard API to drag between views, so I had to implement my own which is fully animated and interactive. Try to dragging tasks to lists on the iPad - the feedback is immediate and you will see the 3D animations when items drop. These are all little details but they take a lot of effort to get right.

- If you run the iPad app, you will see that it has a completely interactive tutorial. We spend significant amount of time on that, to make it feel super nice and provide the best way to acquaint users with the interface. Again, anyone who cares about details will tell you that the final 10% take a considerable amount of the total time.

- You're also not seeing a significant amount of research and testing. We had experimental interactions that never made it to production. We've made many prototypes throughout the life of the app, none of them publicly available. We learn a lot from these but that's all invisible to our users.

The above points just scratch the surface on the work that was needed to bring Clear to the iPad. I hope it illustrates the point that it's not a simple job of just recompiling.

← PreviousPage 4 of 4