I want to use three20's TTThumbsView & TTPhotoView to build something like Grindr. Instead of (1) subclassing these classes, I'd prefer to (2) analyze them and build take the necessary code & tweak it accordingly. Why? Because my concern is that option (1) will increase the download size of our app & slow it's performance. (Another thing I want to ues is three20's caching, but I don't really understand it because right now we're using Core Data. How does three20 caching work?)
If I'm only using part of three20, when I compile the app, will it also build in the other code that's not being used? Or will it leave it out? If it leaves it out, then that gives me some hope that maybe option (1) won't make the app a lot larger to download.
In terms of performance (speed). Will three20 run fast? There's a lot of subclassing going on. Doesn't that slow things down? The Facebook iPhone app seems pretty good, and that uses three20, right?
What do you think?
Thanks!
Matt
There's quite a few questions here:
Analysing the three20 code: People often try this, but it seems not many people have success with this. Here's an example of one project that gave up: http://github.com/enormego/PhotoViewer (the history is all in git I believe, so you can go back and see what they achieved before they gave up)
Three20 caching: can't help you there; you might want to ask a separate question about that. Generally it's best to only ask one question per question :)
Size of three20: I believe these views have a lot of dependencies other three20 classes, so whilst only the necessary code is built in, that's actually quite a lot of code.
Speed: Yep, it's what the iphone facebook uses, so you should be fine there. In fact that's the main reason people use three20 - it's very difficult to do a fast, good photoviewer, and three20 is the only open source project I'm aware of that has done it.
Related
I need to develop a client application for a site that host and show images.
The client should shows all new images and take a picture and send to website.
I'm a newbie to iOS programming so this question:
In an application alike this, there's the need to create a Model? I ask this question because I think I can do anything with Controllers but maybe I'm wrong.
While there is never a "need" to create a model, it's good form to do so, and it'll likely save you many headaches down the road. Ultimately, you're going to need to deal with images and websites and how they interact with each other and whatnot, so you might as well do it in models from the outset.
I'm as guilty as anyone for wanting to see immediate results and hack together something quickly, but I always regret it later on, and have to spend more time than I'd expected refactoring it into something usable.
The MVC pattern as used in Mac OS and iOS applications isn't something best explained in an answer box but here's a link to some simple documentation about the roles of Models, Views and Controllers
Some good basic introductions to start with, which also have relevant links to the next stage of documentation are:
Cocoa Core Competencies
Cocoa Application Competencies for iOS
Read these excellent articles first. You'll learn better and faster if you do some basic reading.
I'm just creating my first project in Cappuccino without a lot of success. Looking for examples and documentation, there isn't a lot out there...
The differences from real Objective-C and the lack of libraries mean you cannot easily convert code directly from say a OS X desktop app. to Cappuccino.
As an experienced PHP, and RoR developer (besides the Java, VB, VC.....) I can't quite see the logic of abstracting/wrapping JavaScript in another language.
Any comments or suggestions for a more efficient framework?
If you have specific questions that can't be answered in the IRC channel (it just depends who's in there) you should post the question to the mailing list and someone will help you.
As for why you would want objective-j I'll point you to this blog post which contains several reasons why objj exists. http://cappuccino.org/discuss/2008/12/08/on-leaky-abstractions-and-objective-j/
I have converted several Objective-C classes nearly line for line into Cappuccino. Mostly it's a matter of changing NS to CP and changing pointers to non-pointers. What differences between Objective-C are you thinking about?
Indeed you could even take your existing XIB files from your Cocoa desktop app and convert them to Cappuccino using nib2cib. Here's how to work with Interface Builder using XCode 4: http://groups.google.com/group/objectivej/browse_thread/thread/786331dbcbc9c7b1.
I also recently started a project in Cappuccino and what I could say is: definitely it worth it. I'm used with javascript as well as (but less) cocoa (which helps). In short, the javascript is very fine, probably one of my favorite language because of it's flexibility. But it does lack of structure. For example there is no class in javascript and you have to trick with object inheritances, prototyping and so on. By using the Cappuccino framework at first you have all those very convenient facilities but also you do not reinvent the wheel each time. This is cross browser standardized so that you only have to focus on your code and not on the IE whims. As it is based on cocoa, cappuccino also comes with the MCV scheme in mind which again let you structure your project in an efficient way.
Ok you probably already know all this. What I could tell now about the learning curve is that I had to find the right tools to really get in. i.e.Installing cappuccino (sudo Starter/bootstrap.sh), creating a simple nib application (capp gen -t NibApplication YourApplication) then using xcodecapp-cocoa to listen, convert and open the project into Xcode what probably the hardest part of this learning curve to me…
Now if you are still not satisfied with cappuccino, I can only suggest to look at another framework like http://javascriptmvc.com/ , www.grooveshark.com used it to do their amazing webapp-site but I personally would never go back to html for building a web application. (conversely I would not use cappuccino for a webpage neither, "The cobbler should stick to his last")
I began writing cappuccino a couple of months ago and had a lot of difficulty when it came to setting up the proper development environment. I also had numerous problems writing working client server code that worked well.
I eventually broke through and started doing better with Cappuccino and then I found these video tutorials that would have helped so much earlier on. They did cost me $30 but I really wished I had watched these at the start because they help you get set up properly as a developer and get you started on the right learning curve.
I actually discovered these videos when using Cappuccinos IRC which has helped me a couple of times when I couldn't find an answer.
Info on the videos here: http://cappuccino.org/discuss/2011/10/19/cappuccino-training-course-ideveloper-tv/
I need to implement cover flow in our iphone app. Our app need to display images from server. We already have webservice to return list of pictures. Number of pictures will grow over time to 1000s of pictures.
Here I need your help to decide which library is good for my situation without any memory issues. and any experiences you have in similar scenario.
I appreciate any input you can give in my scenario
Here is list of libraries in mind.
tapkul
OpenFlow
FlowCover
Please help deciding which library to use and
I can't help on the issue of which library is best to use, but be aware apps have been rejected in the past for doing including cover-flow like features.
If you really must include cover-flow functionality, I'd recommend you write a little test app with each of the available libraries and see which fits your requirements the best. You might also try memory profiling the different solutions and see which consumes the least resources.
There probably isn't a "one size fits all" answer to this question, as everyone's requirements will be slightly different.
Are there any guidelines on pitfalls to avoid while developing iPhone applications?
Sure, thousands. The same is true for any software development. Unfortunately, the easiest way to enumerate them is to write them down on a sheet of paper while waiting for a friendly soul to release you from the one you just fell into.
However:
Don't try to reinvent the wheel. The iPhone API is very complete -- you just have to LOOK for the facility you need. Things are NOT always implemented the way you would expect. Read the guides, carefully. Look at the tutorials and analyze how they work. (Try changing a line here or there in the tutorial to see what difference the change makes.) The single biggest mistake I have made in 1 year of iPhone development is not trying hard enough to find the iPhone way of doing something.
Don't ignore memory management; master it early and often. Use the Object Allocation and Leaks tools in Instruments to check for memory leaks frequently. I'd recommend checking after you complete each feature or view; more often than that if you keep finding bugs. Eventually you may understand it so well you can stop doing this.
Don't just use the default build settings. Play around with them to understand what they do. Figure out certification and distribution. GET INTO THE DEVELOPER PROGRAM QUICKLY -- it can take a while to push through that pipeline. [ AND when you get that notification that you need to renew, get it on instantly -- there have been problems with that process. ]
Don't neglect to read the Human Interface Guidelines (HIG) carefully. If they say not to do something -- DON'T DO IT. Apple will reject applications that misuse their iconography.
Don't stint on marketing. Yes, the App Store puts your app in front of millions of people... In theory. But the odds of getting front-paged are slim. There are a lot of great apps on the App Store that haven't sold much because no one knows about them.
Don't rest on your laurels. If a new technology comes out, find out if it makes your job easier; if it does, take the time to learn it. Personal example: I'm just now trying to switch from SQLite-based data management to Core Data, because I was in a hurry at the time I started my most recent project; now I wish I had slowed down and thought about it.
Don't go into your design thinking (for example) "How do I implement my concept with a table view?" It's true that table views are natural for many informational and utility applications, but don't be constrained. Instead, think about what users will want to be able to do, how you can make it easier for them -- put things together that will be used together, etc. If you've never explored the concept of Use Cases, read up on them.
Don't hesitate to build composite views. Many of the questions I have seen here on Stack Overflow have to do with putting a toolbar at the top of a table, or having an image in the background of a text field. I understand the desire to do things the easy way, and as I state in #1 above, if there is an easy way, use it. But in many cases the solution is just to layer a couple of views with appropriate placement and transparency.
Think about what might be Apple-approved from the start.
App Rejected is one of several useful sites to help understand Apple's mostly undocumented standards. (One more.) (A previous question on app store rejection reasons.)
A few quick examples:
Using a UIWebView can get your app a 17+ rating.
Coding with an undocumented/private API = rejected
Version number < 1.0 might= rejected
Not enough feedback about network success/fail = rejected
Too much network use = rejected
Clearly limited free version vs full version = rejected
The word 'iPhone' in the app name = rejected
The above links contain many more examples, and more details about those examples.
Don't neglect the programming guides. While the documentation is quite extensive, the programming guides contains a veritable trove of useful tips and "insider" information that simply cannot be gleaned from reading method definitions. I spend just as much time reading the guides for a technology (say, Core Data) as I do actually implementing it.
Don't assume you know what a method does. If you have any degree of doubt about the functionality of a method, it is well worth your time to go look it up in the documentation to verify.
Wonderful examples from #Amagrammer above.
I would love to add that the first place to start is iPhone development is Photoshop. This is still the best advice I can give to anyone who is starting out. I now use OmniGraffle because it has awesome stencil templates.
What I find is that even for super simple app's, draw up your prototype and look for usability issues and work flow issues. It is 100x quicker to redraw your app than re-code it. I have fallen into this trap numerous times and now actually draw up some pretty simple functionality to see what it will look and feel like.
This advice will save you 10s maybe even 100s of hours in hopefully getting your app right first time and getting you to think through what the issues are. Throwing away code sucks and I have done it not because the code was bad but because it made the usability or solution worse. I think the best of us end up throwing code away and prototyping your design definitely will help in having to RTFM for something you did not have to build in the first place.
If you don't have an great designer, and can't do great design by yourself, then don't even start iPhone app development. This rule only applies if you want/need to make money with your apps.
I've been evaluating a framework that on paper looks great. The problem is that the sample code is incomplete and of poor quality. The supplied reference implementations are for the most part not meant to be used (so they can be considered as sample code as well) and have only succeeded at confusing me.
I know that it's common for things to look better on paper, but my experience with the sample code is turning me away from further investigation.
Do you let poor code samples change your judgment of frameworks/libraries? So far my experience has been similar to the "resume effect": if someone doesn't put the effort into spell checking their resume, they probably won't get the job...
For me, it does. I tend to want to avoid libraries where the code samples are incomplete. If the library is open source, I will overlook it, since I can directly look at the code and see if the library's internals are reasonable, and I know that, if there is a problem someday, I could (if I had to) fix it.
If the library is commercial, and their samples and/or documentation is poor, I look elsewhere. I just see it as risk management - poor samples make me fear the quality of the library in general.
No matter how good something is on paper or in theory, it can still be crap when programmed.
I think this is a valid reason to turn away from and evaluate other libraries. As a potential user of a library a lack of documentation and/or bad code samples gives the impression that the library is not yet mature enough for use by third parties. In time it may well gain the missing pieces but until then I think its reasonable to look elsewhere.
I was recently evaluating the multitude of blogging applications that people have uploaded to github.com I quickly skipped ones that no documentation as they obviously weren't ready for others to use. The ones that remained at the end had a good README with info on how to get the app up and running as well as an online example of the code running.
Poor code samples combined with poor documentation will make me turn away from a library unless there is a compelling reason to use it. However, a library that has either good code samples or good documentation is usually worth using. (Assuming that the library itself otherwise meets my needs.)
If I can't find good examples (and/or documentation) illustrating how to use the library, I'm definitely less likely to use it - just as a practical matter, it'll be harder for me to figure out how. But I don't care what the code that implements the library itself looks like. I don't think I'd choose one library/framework over another just because the developers of the one have shown an ability to write cleaner code (which is what I understand the "resume effect" to mean).
Lack of documentation and examples makes me a whole lot less likely to use that particular library. It's not worth my time testing and trying to figure out how a black box works if there are alternate solutions to the problem out there.
Yes, definitely. Every library should come with a simple example using program and a CLI interface (for very simple libraries with <3 methods and <10 hooks, one example should suffice).
And why does your framework "look great" if it's so hard to use that even the original coders make mistakes using it?
It certainly matters to me. Evidence of sloppy/incomplete coding and poor communication decreases my confidence that the actual implementation code is stable and robust.
Myself yes, but there must be people out there who aren't turned off by this otherwise there are plenty of open source projects that would have died a long long time ago.