back to article Google kills off app maker

Google's App Inventor could be revived as an open source platform in order to let students click their way to Android applications without having to muck about learning stuff. App Inventor was run by Google Labs, the research operation currently being shut down as part of the chocolate factory's streamlining focus. But in a …

COMMENTS

This topic is closed for new posts.
  1. Turtle

    Imagine that!

    Ah, something heavily promoted by Google turns out to be a complete non-starter in practice.

    How often does *that* happen?

    Actually, now that I think about it, it seems to happen quite a bit.

    Well I guess that it's happened again.

    Imagine that!

    1. Anonymous Coward
      FAIL

      Perhaps

      But at least they got off their asses and tried, unlike some companies that are either too scared to fail or don't spend money on R&D at all.

    2. Anonymous Coward
      Boffin

      It's called R&D

      All R&D involves of trial and error. That's why it's called research.

      Actually I used App Inventor. It worked for simple things, but was not that simple when it came to making full-fledged apps. It was a very Beta product, so that's fine.

  2. Anonymous Coward
    Anonymous Coward

    I don't think it's the speed, really.

    There's some atrociously slow, but still quite popular, non-drag-and-drop "app makers" out there. The clickibunti approach is perhaps a natural extension from a developers' PoV, but apparently not that much so from the purported audience's PoV. Perhaps it's because GUIs exist to "be intuitive", aka to forego thinking, and the act of programming is the act of pouring thought into machine instructions, maybe?

    We could do with more thinking, really. The non-thinking part we already have pretty well automated.

  3. Anonymous Coward
    Anonymous Coward

    Google App maker

    wouldn't give me any APIs to the bits of the hardware I wanted (microphone) (SD card) it was easy to write a demo app on the accessible bits, just useless for my needs. so I can delete the eclipse framework now and get some gigs back on my eeePC!

    1. M Gale

      Eclipse framework?

      Uhm, App Inventor is the IDE. You use it instead of Eclipse, not alongside it.

      As for how good I found it, well it seems like a very very good idea, ruined by having it as a "Cloud" app tied to Google. Go to URL.. wait for page to load... log in.. wait for page to load... open new project.. wait for page to load.. open blocks editor.. click to confirm I want to open a jnlp file.. wait to download.. wait some more to download.. click to confirm I want to run the jnlp file.. click to confirm I want to continue waiting for the jnlp file to run... wait for the jnlp file to run... oh heck, finally I can make the HelloPurr example. Heaven forbid I should try this on a train or something with intermittent Internet availability.

      If it ran locally on my own Apache2 (or could just be installed with its own localhost or LAN-only web server) then this could be awesome, and quick. As-is, it feels like pulling teeth with a monkey wrench. Slow, painful, and using all the wrong tools. Maybe the ditching and moving to open-source will mean the above can happen and we get a RAD environment that actually runs rapidly.

  4. Peter 26
    Thumb Up

    Good it was terrible

    I wrote an app in AppInventor and it was painful to the extreme. 20 lines of code in java for snake turned into a week of work and something looking like a crazy A1 poster size spaghetti map. It also was incredibly slow despite doing things not that complicated.

    I suppose it sets you on the path of coding android and lets you actually produce something straight away which you can run on your phone, making you eager for more. But it is useless for developing any app you would be proud to go public with.

    TL;DR AppInventor is a bit of fun, nobody should be using it for making real apps.

  5. dogged
    WTF?

    Alternate Sub-headline

    "David Pogue cannot tie own shoelaces".

    I mean, seriously. This is point&drool stuff. There's a special needs school down the road from me where the kids made an app using this as part of a charity challenge.

  6. Eddie Edwards
    Facepalm

    Runs slow?

    "until the world remembers that natively-developed code runs faster"

    Er, 1985 called and want their reason not to use productivity tools back.

    In fact there is absolutely no reason in principle that a graphical tool couldn't produce code as fast as native C or even native assembler. The reason people don't use graphical tools, so they never mature to that level, is that a few lines of C translate into an A3-sized diagram, and no-one wants to wrangle an A3-sized diagram. Which is sad, because once you get above the low level to, say, integration of heterogeneous systems over a network, the diagramming approach should win out by a mile.

    This is the same reason, incidentally, that most chip makers write VHDL or Verilog instead of making football-pitch-sized schematics. (Ironically, in that case, the people that do use football-pitch-sized schematics do so to get extra speed.)

    1. Dagg Silver badge
      FAIL

      Maintenance

      The figures I had heard is 80% of the cost of software is maintenance. These GUI tools might be able to create something but are absolute rubbish at helping to maintain the resulting code.

      Each time you want to make a small change you end up having to regenerate and rebuild and then because so much stuff has been touched you end up having to rerun all the existing module tests, regression tests, integration tests, system tests etc.

      Meanwhile in a native app someone has made the small change and tested just the area that was changed.

  7. Mage Silver badge
    Coat

    Anyone here old enough?

    Remember the hype about "The Last One"?

    http://www.tebbo.com/archive/wf8104.htm

    Or Smalltalk

    Or C++

    Or Java

    Or .Net

    Or PHP

    Mine's the one with Z80 Instruction set in the pocket

    1. ThomH

      To be fair...

      ... all of those except The Last One (if you give Smalltalk a pass as the basis of Objective-C) have been used for a bunch of useful, productive applications and are still being used today. The hype is always a bit ridiculous, but useful ideas have emerged.

    2. sabba
      Facepalm

      Ahhh!!

      <<sighs nostagically>>the good old Zilog Z80 and Mostek 6502!! Now those were the days. Handcrafting code down at the instruction level. I remember a time when i knew the numeric codes for the entire Z80 and 6502 instruction set. Then I discovered girls and alcohol...where did it all go wrong!?!? :-)

      1. T.a.f.T.
        Joke

        C

        > where did it all go wrong

        You didn't learn C which gives a reasonable level of control but leaves enough time at the end of the day for girls & alcohol (just don't say any more about your job than I am a programmer for X(applciation) and your odds of getting laid stay about par with the average man on the street).

  8. Eponymous Cowherd

    Drag and click. Can be good, but usually isn't

    Scratch (http://scratch.mit.edu/), which appears to be the basis and inspiration for App Inventor, is really rather good. But, then, its an introductory teaching language. Trying to scale up a fun and inspirational teaching tool into something for creating usable apps is, really, a non-starter. I've tried App Inventor and really cannot get on with it. Its far easier and quicker to write apps in Java using the Eclipse plugin.

    Speaking of Java. Anyone remember "Visual Age" for Java? They were using it at EDS around 2000. That was a drag-n-click environment, and was downright nasty. You "wired" objects together. It produced reams of spaghetti code that you had to edit to do anything useful.

    The King of drag-n-click languages has to be NI's LabView "G". Now that is good. Really good, and enormous fun to use. Of course, its "dataflow programming" paradigm is well suited to its primary task of running test benches.

  9. BitDr

    Drag and drop programming

    Let's not forget Amiga Vision. That one followed the flowchart paradigm and required thinking, but was very useful (I once wrote a video store info kiosk with a light-pen for user interaction with the GUI).

  10. Anonymous Coward
    Coat

    Releasing App Inventor as open source

    "Releasing App Inventor as open source will require a careful examination of the code to establish whether any bits are owned by someone else, but restricting it to educational use might mitigate against the patents on which the platform likely infringes."

    Really? Google checking if the code belongs to someone else (like Oracle perhaps)?

    1. Anonymous Coward
      Anonymous Coward

      Not just the code

      But all the possible ideas too, like "having a jigsaw piece that performs an action" which I'm sure someone has a patent for somewhere and is just waiting.

  11. kPkPkP
    Megaphone

    MIT Center for Mobile Learning to study & extend App Inventor for Android

    from: http://googleresearch.blogspot.com/

    "MIT and Google have a long-standing relationship based on mutual interests in education and technology. Today, we took another step forward in our shared goals with the establishment of the MIT Center for Mobile Learning, which will strive to transform learning and education through innovation in mobile computing. The new center will be actively engaged in studying and extending App Inventor for Android, which Google recently announced it will be open sourcing."

This topic is closed for new posts.

Other stories you might like