Teoti Graphix, LLC

Hi,

Well I have made it 3 weeks in a row with the flex-dev summaries and then yesterday a nice winter sickness. It’s funny how being in an altered state can get you thinking about things.

For now I am putting these summaries on hold since well, in my opinion Apache Flex is at a stall right now. JIRA migration and full module support is definitely proving to be a slow down for any real progress. I have some other projects that need attention and talking about more or less gossip on the list is not what I call fun.

I hope the issues with Adobe get cleared up sooner than later. I was ready to get going on this project but, there is really not much you can do right now unless you can afford to burn money and time chatting like at a bus stop.

I will be back on the flex-dev airwaves as soon as there is something to talk about.

Mike

Monday always has a bit of a bite. Sentiments of Monday followed this with various subjects revolving around inpatients. I guess it’s only natural when your future hangs in the balance of belief.

Logo final round

Well, Doug posted a thread yesterday for the final round of the logo contest. This will decide the winner and communities favorite.

For information on voting see;

http://incubator.apache.org/flex/logo-contest.html

Component cloning

This was an out of the blue request, DisplayObject cloning. I have to admit I have never really needed this in the apps I have made. Alex mentioned and API could be added to create such a feature but the, what are the use cases?

From my perspective, if you can’t create a utility from the outside that calls public API on a component to clone it OR holds state that initializes the component from say an application model, you are trying to do to much.

I don’t think this will ever be an option in the framework. There really are already ways to accomplish this as stated.

Component Behavior Pattern

This topic has been flying around not just in the mailing list but on a couple blogs as well. I am no stranger to behaviors, fact is back in 2008 I wrote a complete framework from Flex 3 using them as a prototype experiment. Didn’t really get any exposure, I did get four favorite stars in google code for it!

So anyway, this subject really relates to the original thread I started on the list about Modularity and Interfaces. As Alex keeps pointing out, we live in a constrained environment(Flash Player), rules have to be bent and broken all the time to provide optimum performance.

There are draw backs to composition as Alex points out. A major consideration is the line between granularity and performance. It’s noted that although breaking a class up into parts allows flexibility, we then incur the cost of calling the proxied methods of the composed object. So here is the problem, there is not a way with the current constraints of the system to actually create a “purist” behavior pattern. Oh the engineers in us wish we could because it would make life 10 times easier, but we have to realize that any design pattern we can come up with is going to have the flaw of overhead in the Flash Player.

State of the project

The state of the project was discussed, what is that you say? State, you know where is it at. What are the future goals, what promises have been made, why are we here. I feel these answer will come in the next couple weeks. The wiki has been created and I think there is an effort to have a page that allows for the community to see the progress of all these questions.

Coding conventions

Coding conventions and best practices was brought again by a new community member, it seems as though that subject has quote “been beat to a pulp”. So a community member was kind enough to use markmail links to guide the new member to the 100′s of threads written on the subject.

Ah, how sweet is that.

Tagging on the list

Tagging for list threads was also mentioned for adding more clarity to a subject line of a post. IE [SomeCategory], explains where the subject belongs in a namespace fashion.

Folks, I have had enough experience with forums and mailing lists to say, whatever order you try and achieve will never work. Mainly due to the influx of new members and the fact it’s community driven from all over the world.

Moral of the story is, create email folders like flex-dev, flex-user, flex-commit, setup email filters to push anything from the said emails into the folders. As an added bonus, use an email client that can sort mailing lists by thread when you are looking at them. You good to go using this pattern.

Note; The only flaw with this is when members start new subjects in a reply to another subject. Then threading goes to well you know.

Conclusion

Monday, its definitely not Friday.

Thanks for reading,
Mike

Well the weekend was probably the summation of a very active 2 weeks on the list. The main highlights were the SDK being committed without automation and required modules. Moving ahead in conversation today, various community members are voicing their frustrations at Adobe for legal issues regarding the above missing modules.

I have to admit, it’s hard to get on the road to the future when your still looking for the other shoe which is happening right now. On a positive note, what will be will be and the more time you spend talking about what isn’t the less time you talk about what is and will be.

Enough philosophy, my point is the future of a lot of developers ride on the next two months and what will happen to the Apache Flex project as a whole. I can’t stress enough that a project moving from a commercial entity to opensource like Apache is BOUND to have legal problems in the beginning. It’s up to us right now to fix and create a base structure for the time when these issues become non existent.

Adobe and legal statements

It has been asked on the list that Apache Flex get solid answers from Adobe about the missing modules and a time frame when these things will be resolved. Is this actually possible? Nope, and the more time debating who screwed up or where the answers come from just fragments the community further.

Community patches

Here is another area where patients wins. JIRA is the place for the community to place patches for the SDK. Yup, JIRA has not had the Adobe issues migrated yet. We need to wait, well, just wait.

Apache Flex versioning

The most talked about subject since the logo. What seems to be the consensus? Using years is prone to debacle, sticking with Adobe’s versioning scheme sounds like the consensus, so there you go.

Issues with the build

Yes, more issues but the build was not completely tested. The README.txt was tested and does work, so here again to soon to expect inflated tires.

Conclusion

Have you ever tried moving from one house to another? Major stress, disorganization, missing possessions, staled agreements, etc. This is the same dynamic happening with Apache Flex. If you can see the future and that a larger home will benefit, patients will always be the best remedy for issues that are utterly out of ones control.

Thanks for reading,
Mike

Short and sweet today. The code is in, many different ways of building it are surfacing. From the command line, maven and Flash Builder.

I will get a more comprehensive list this week as the code settles in a bit more.

Conclusion

The calm after the storm. Voting is over, code is in. This next week should be more about the future plans of the project now. What is going to happen to what and where?

Mike

Hi,

I just have to post this image since well, it was a first for me actually. I have never once actually built the SDK using the build file when it was opensource at Adobe. I was just another user downloading new versions and updates.

Well, I am a visual person by choice, the command line I try to stay away from at all costs. So I thought for those that haven’t touched the command line in a while would like to know that it is possible to build this monster with Flash Builder. I don’t know if you can just use Eclipse with ANT, I haven’t tried that but I’m guessing there are some settings in FlashBuilder that allowed me to build it so easily.

Basically you need to do this;

• Download the Open Source 4.6 SDK – http://opensource.adobe.com/wiki/display/flexsdk/Download+Flex+4.6
• Choose a location for the new SDK and extract the archive.
• Open the extracted folder’s frameworks directory and delete all contents.
• Checkout the http://svn.apache.org/repos/asf/incubator/flex/whiteboard/frameworks into the emptied frameworks folder. (this took 15 minutes on a fast connection for me checking out the whole trunk)
• Open Flash Builder and create a new default Project (just a simple project no natures)
• Create a new folder in the project, Click Advanced and “Link to alternate location”
• Navigate to the frameworks directory you just created and select it for the folder location.
• Open the frameworks folder in the new linked folder.
• Right click on the build_framework.xml and select Run As Ant Build

You should see ANT’s gears turning in the console. The build took my laptop 5 minutes to complete after it downloaded the air and player dependencies.

You then create a new SDK pointing to the new folder that was extracted earlier(the parent of the frameworks directory). Then create a new Flex Project using that new Apache Flex 2012 SDK!

And voila, we have my first hello world app with Apache Flex binary!

Note: I used the framework directory link for simplicity since all I was dealing with was the root of the frameworks directory.

There is currently a bug that I want to fix in the build file but I need an answer on which flash player the SDK is using. As of today to get this working you need to change the player/11 directory to player 11.1. This will be fixed soon.

There is a difference in player numbers between the build.xml and build_framework.xml files that needs to be fixed.

Let me know if this doesn’t work for you.

Mike

Well folks, the first drop of the Apache Flex SDK is now in the whiteboard, it’s your job to build it and test in your applications/IDEs. Lets see if anything broke so we can get a first release out sooner than later.

SVN commit r1234192

This will be the commit that has established Apache Flex as a community first framework. Carol managed to commit the framework to the whiteboard/frameworks folder in SVN without SVN with history for now.

http://svn.apache.org/repos/asf/incubator/flex/whiteboard/frameworks

The framework does not have the compiler, mustella test suite and automation yet. These modules have yet to pass through Adobe legal and will be contributed later.

JIRA issue migration is the other important factor that is still being worked on as well.

Logo contest

The logo contest is at it’s final day of voting so if you haven’t voted yet, get on the list and submit your vote.

More information for the rules of voting here;

http://markmail.org/message/zfr64u4guwvlumcc

Button.buttonMode = true

Once the source was committed, Omar suggested a first commit and that had to do with the spark Button’s buttonMode property set to true.

This then started a conversation about being careful about version compatibility and breaking someones app or altering the expected user experience.

Alex then suggested creating a style for that change. This way, you keep the default backwards compatibility and also allow developers to use this hand, pointer custom cursor setup in a style theme.

As the conversation proceeded with community members chiming in, it was also noted that on top of the style you could then customize the cursor with Embed or ClassReference.

s|Button
{
    cursor: ClassReference("com.project.CustomCursor");
}

OR

s|Button
{
    cursor: hand;
}

I’m just using this first example as proof of the power in community decisions and brainstorming. Whether this change is committed is up to the community but, one thing is for sure, “It’s up to the community!“.

Faster GroupingCollection

Ryan Frishberg added a refactoring to the GroupingCollection2 in his whiteboard. He says that it’s 1.5-2x faster than GroupingCollection2 and using summary rows tends to be around 50x faster GroupingCollection2.

If community members are interested, he said we could talk about adding it to the main framework once testing is up and running.

Source; http://svn.apache.org/repos/asf/incubator/flex/whiteboard/frishy/FastGroupingCollection/

Apache Flex SDK versioning

The sdk versioning is something that is going to need to be figured out and there are really two options.

- Keep using the old Adobe 4.x tradition
- Break from Adobe and call it Apache Flex 2012.x

The last option uses the year as the main version number and also establishes a pseudo version 1.0 for Apache Flex. This option was suggested by Alex and I will quote him;

I just got confirmation that FlashBuilder will not be using years, so my proposal is that Apache Flex uses years. The first release, even if it is just approximate parity with Adobe Flex 4.6 would be called “Apache Flex 2012″. Any other release cut this year would be a point release (“Apache Flex 2012.1

Your duty as an Apache Flex community member

Dear development community, this is the best way to get involved in the Apache Flex project right now, build it!

I haven’t actually done this yet but, you can best bet I will have a visual guide up sooner than later showing people how to do it.

For now, there is the all powerful README.txt in the root of the frameworks folder.

See; http://svn.apache.org/repos/asf/incubator/flex/whiteboard/frameworks/README.txt

For now, it will take some one that has some experience building with ANT to get it going. I have on my list of todos to create a visual tutorial to help other community members get on their feet and build with ANT using the README.txt instructions.

The more community members we can get to build this and report back to the flex-dev list about problems, the faster we can get Apache Flex up on it’s feet and get that first release out to show our supporters how solid our development community is.

Conclusion

I have to admit I was surprised to see the code. That was a great surprise though! Now the adventure begins and we can work together to get Apache Flex stable, setup and release Apache Flex 2012!

Thanks for reading,
Mike

The week flew by with the logo contest and voting, huge community involvement, the creation of the flex-users mailing list and some rock solid information on the Falcon compiler.

SVN code commit

Well unfortunately the Flex 4.x source code has not made it’s way into the repository yet. The problem now is the Adobe donated code is staged but it includes a lot of Adobe’s other projects as well that need to be filtered out before Carol is able to hand the source code off to Apache Infrastructure to do the final dump (keeping the history is the major problem along with the sheer size she is dealing with to filter).

So we are keeping our fingers crossed for next week! Again I will say, you have to give it up for Carol and Alex. The amount of time they have put into this source code with legal, worrying about migrating Adobe JIRA bugs, etc. I am sure they are ready for a vacation.

Logo contest voting

The logo contest voting is happening as we read so if you haven’t voted for the new Apache Flex logo, give it a try! The vote lasts until 01-21-2012 around 9PM GMT I think.

More information for the rules of voting here;

http://markmail.org/message/zfr64u4guwvlumcc

and you can see the Flex-Dev 15 for more information as well.

Falcon compiler update

Well this subject was so exciting for me I couldn’t even wait posting about it, see yesterdays blog post.

The short story; Gordon Smith from the Adobe Falcon team was kind enough to update the flex-dev list with enough information to get any parser/compiler freak excited.

Features

- designed from the ground up with cross compiling in mind
- designed to support multiple compiling targets in multiple projects in a workspace
- multiple threads to compile multiple files at the same time
- In memory workspace/project data structures give highlevel symbol table of class/interface/functions etc
- Solid Document Object Model for read/write on the parse tree
- understands .as, .mxml, .css and .properties files
- parse trees for AS consists of about 100 classes such as LiteralNode, BinaryOperatorNode, ClassNode
- MXML parse tree looks the same with nodes such as MXMLDocumentNode, MXMLInstanceNode, MXMLScriptNode
- uses 3rd party grammars to create the various lexers, parsers and code generators with ANTLR, JFlex and JBurg parser generators
- MXML is compiled directly to ABC, not AS source code or AS parse tree

From a tool developers perspective this is such great news. These specs given also allow some prediction that a whole new ecosystem of tools can emerge because the Code Model and DOM is not connected to any IDE!

I’m sure I will be posting more on this subject, stay tuned.

See; ApacheFlex :: Falcon compiler update

Trace & Log

A community member that is well know in the logging community was testing the water to see if there was interest in compile time injection of file and line location of log calls (plus a bit more but I’m not a logging guru).

- Granulate logs to methods or files.
- Less coding for the user. Granulate without currently needed boilerplate code.
- More verbose. Show where the particular log statements have been originating from.

Quote from Martin;

The granularity is important if someone has long files, I am thinking of the UIComponent example right now.

The verbosity (line number and location) is very important for bug tracking or code functionality understanding.

The simplicity is very important for developers who dislike to write a lot of boilerplate code.

Apache Flex wiki

The wiki came up again and looks like there are some community members that could put this to use sooner than later (myself included).

I could see this possibly turning into another group “learning experience” if we do not agree from the start how to structure “places” on the wiki.

I do love the versatility of wikis but, if they are not at least thought out a little bit, navigating them can be a pain in the WIKI.

Conclusion

It’s been a fun week, I think about the only thing that didn’t happen this week that the community thought would happen was the source code commit(but we all understand why).

If you were to skim these summaries for the week you would find the following highlights;

• Logos, logos … LOGOS!
• Apache Flex Versioning
• Flex Charts
• DataGrid Performance
• DropDownList working on Mobile
• Spoon Community Call Feb 2nd
• Falcon compiler source code / Falcon architecture
• Apache Flex JIRA
• Method Overloading
• Source Code Commit
• Lightweight Event handling
• Dependency on Adobe
• Line size determines code quality

Thanks for reading,
Mike

Logos, logos … LOGOS!

By the time these last two weeks are over, I don’t think I will need to say or think about the word logo for awhile. Kind of like the beginning of a box of chocolate, by the end it’s more about finishing the box than enjoying the last ones.

I’m glad this logo “fiasco” happened, for without the mighty grace of the word LOGO there would be 1000′s of people that have no idea what Apache Flex is right now. From the standpoint of advertisement, I think we hit the mark as community members.

For factual information, the LOGO Vote is happening. Ah yes, community votes are coming in right now.

Markmail

Markmail Logo voting Instructions

Apache

Apache Mail Logo voting Instructions

Please vote via mailing list (flex-dev@incubator.apache.org on Topic “[VOTE] Logo contest“).

For now, be prepared to see the subject line “Re: [VOTE] Logo contest” many times if you are subscribed to the flex-dev mailing list.

Spoon

Spoon’s presence and mission was an interesting topic. I had mentioned the open call that Jonathan brought up on the list yesterday. Since this is a world wide project, there seemed to be an understanding in misunderstanding.

Michael Labriola from Spoon also tried to smooth out the wrinkles of misunderstanding. Check out the rest of that thread if you’re interested.

Mike’s outlined Reply

Flex 3.x SDK Support

The topic of the Flex 3 SDK support is something that has come up again. Community members from the enterprise have chimed in stating that a lot of major enterprise flex applications are still using the 3.x SDK. There was an initial question as to whether the SDK will be supported in Apache Flex. This sounds like a no brainer but, it wasn’t clear when this code base was going to be contributed.

Long story short, Alex and Carol seem to be having a little bit harder time getting all the legal preparations and clearances for the Flex 4.x SDK that originally expected. I’m sure once the 4.x version is finally committed to SVN, 3.x will be the next hurdle to jump and will eventually be in SVN as well.

Alex mentions;

Right now, I think there will be a window of time in April and May when I was going to make contributions to 4.6. (After that, I’m hoping to start prepare Falcon for Apache). If we hear enough requests for 3.x vs 4.6 work, I would consider trying to get 3.x donated during that time. It is pretty much up to the community to decide.

Sounds like a JIRA issue and votes to get that going sooner than later.

Apache Flex Versioning

Apache Flex versioning has been an ongoing topic and is also one of the initial issues filed in JIRA. There is debate on “how” to version the next release of the Flex SDK relative to the Adobe Flex version 4.6.

See the thread titled;

Re: jira task to decide on Apache Flex version number needed

There are to many ideas and noise to get you the clear picture in a paragraph. I’ll post more when a conclusion seems close to being reached.

Conclusion

Numbers, pure numbers show the future of Apache Flex. We have proven that there is an overwhelming support for Apache Flex through the logo contest’s ups and downs. Let’s just say this was the silver lining in the dark cloud dissipating from last November.

Through honest effort and community it’s possible to create movement even if we have to get out and push from behind to jump start perception again.

Facts and figures show the flex-dev mailing list just hit 2500 posts in a little more than 2 weeks!

Traffic statistics cover a total of 10 days.
Current subscribers: 237
Current digest subscribers: 3
Total posts (10 days): 1116
Mean posts per day: 111.60

PS I have had A LOT of community members say they follow these summaries and are not subscribed to the list, I think we are growing a community here but, don’t take my word for it.

Thanks for reading,
Mike

Here is a nice and full summary for 14 days of blogging!

Flex Charts

The current status of the Flex Charting was brought up and whether functionality and performance would improve.

CSS was a major candidate for improvement and was actually asked that CSS be made a priority for the whole Flex 4 SDK. This means add CSS properties that would allow for display adjustments and shorthand like “real” CSS.

I think users want to see the Flex CSS leave the dark ages. It was also noted again that in the spirit of Apache, roll up your sleeves and try and fix what bugs you, then submit it to the community!

It has also been noted that Adobe doesn’t have any new charting code to contribute. Maybe some specs and prototype code but that is all. Looks like it’s up to the community to see where a new charting library magically appears from.

DataGrid Performance

The performance of the DataGrid and the DataGrid itself has been the nemesis of the Flex SDK from day one, no joke. It’s no surprise that this has come up as about the 3rd request from the community to look at.

The DataGrid has a bunch of problems, it’s idea in the real world can do a lot of things. Add on top of the functionality it requires, the component will get the most abuse from jamming huge data sets into it. On top of that mess you then want to edit and draw that data in certain ways based on data state, etc.

The above boils down to a huge mess if the component is to do everything. It sounds like there are some community members that have come up with some factory implementations that might just tame the mighty DataGrid design. Let’s put it this way, if Apache Flex in the next year or so can create a DataGrid that performs better than the last two, major win for the community.

DropDownList working on Mobile

Tink had brought up a simple little fix for the DropDownList to work in mobile applications. Another community member commented that it would be nice to have this component available in mobile because the Spinner is really heavy for just displaying small data sets.

Spoon Community Call Feb 2nd

Jonathan Campos announced that there will be an open call for community members to ask questions about the direction and focus of Spoon.

At the same time Spoon will be opening up the organization to the community.

There is also a Flash/Flex User group tour starting soon as well, you can get more information here;

http://blogs.adobe.com/flex/2012/01/announcing-flex-user-group-2012-tour-north-america-dates.html

I think an interesting question that will be asked is Spoon was founded before the knowledge of the donation to Apache. How does that change the Spoon road-map concerning the Apache way of “if it didn’t happen on the list, it didn’t happen”.

How does Apache feel about an outside project like Spoon?

I’m sure this is what the Spoon community call is all about, so don’t miss it if you have the time. For now they will be posting more information up on their site;

http://spoon.as

Falcon compiler source code / Falcon architecture

Alex updated the list on the subject. It sounds like the compiler might be available in the summer in an unfinished state.

I’m going to quote Alex here as these are pretty heavy statements;

Falcon is intended to serve not only as the compiler for FB, but as the code-model service for FB’s code intelligence.

The version of FB that integrates Falcon will not be designed to use a different version of Falcon for different versions of the SDK. This is different than today, where FB uses he MXMLC from the SDK version for your project.

He goes onto note the implications;

This has some interesting implications. One is that we can’t keep rolling out new versions of Falcon without factoring in backward compatible MXML code-gen for older SDKs. Another is that, if you muck with the language, the code-intelligence in FB will break or at minimum, won’t be able to assist in those new language constructs.

For the first issue, the plan of record was to have Falcon convert MXML constructs to data structures instead of code, and modify the SDK framework to interpret those data structures. That work will likely not be fully complete at the time of donation.

So what does this all mean? It means that the yellow brick road is full of obstacle and unknowns right now. As a development community we need to focus on the Flex SDK until the tin man at least gets his heart.

JIRA

Well JIRA is definitely up now, Bertrand has now added about 7 issues and tracking features. So the ship has launched.

Please read AND follow the warning on the main page;

JIRA issues from https://bugs.adobe.com/jira will be imported later, please DO NOT create duplicate issues for those.

https://issues.apache.org/jira/browse/flex

Logo Contest

Finally, the logo contest which is at it’s end for submissions today. I created short urls with logo numbers on the web site now, so it’s a bit easier to tag entries.

As far as voting, we still have a couple days as far as I know. More on that later.

http://incubator.apache.org/flex/logo-contest.html

Conclusion

Well, this was an action packed post. As you can see the community is chiming in and starting to get their boots muddy with requests and complaints.

JIRA is your ticket to being heard but, hesitate please before you file a bug or feature request because odds are, it was filled in the Adobe JIRA and will be migrated.

Oh yeah, this is #14, two weeks in a row! I will now pat myself on the back for something I never thought I could do, 14 blog posts in a row.

Thanks for reading,
Mike

Monday was an active day of logo entries, method overloading, source code and event handling discussions.

Pretty soon we will be talking about who is tackling what and whether commits are reverted. What bugs to fix and what fat to cut out. How to implement the future and what is priority for the next couple releases to show that there is a capable community ready to take Apache Flex to the next level.

Logo Contest

This should be the final day I talk about the logo content as it winds up after a week of entries. It was really interesting to see all of the entries thus far and how much they varied.

As usual head over to the logo contest page and check out the new entries.

http://incubator.apache.org/flex/logo-contest.html

Method Overloading

The topic of method overloading came up and since there were a whole bunch of threads, I had to rewind back to where this originated from.

Don’t take my word for it but, I am pretty sure this topics beginning came from Alex talking about overloading and the new framework he has been working on(last week). From this standpoint the overloading has to do with backward compatibility to the older Flex SDK.

From there members started talking about changing the compiler to add functionality A, B and C. Debates were expressed and we arrive back at a point blank question about method overloading.

Here are three items to think about;

- descriptive names verses method overloading
- cross-version backward compatibility
- trade off between overloading vs API bloat

Source Code Commit

The source code “should” be committed today and at the very latest tomorrow. All the legal papers have been signed off by Adobe and Alex is now just waiting to get the SVN dump file from the Adobe servers to hand over the to ASF Infrastructure to dump into Apache’s SVN server.

Exciting to say the least, focusing on issues that are not related to code in a code related project has been backwards for sure but, shows that the Apache Flex community knows how to keep the fire lit until more wood arrives.

I think the fire burning for the last two weeks was purely from ideas and communications. We can now take that energy and put it into a first time release for Apache Flex.

Lightweight Event handling

The question was brought up about interest in a lightweight event model for a branch of the SDK targeted for mobile, high performance small foot print type applications.

The AS3Signals framework was mentioned as an option. There are as always two sides, one that sees the signal framework as a performance enhancement, the other that does not want things added to the core that can be included as an external library and implemented on a per application basis.

Whatever the case may be, this seems to be a recurring debate on how the Flex SDK is going to evolve and what will be included in the core. Only time will tell.

These issues do bring us back to an original thread about the modularity of the Flex framework. Whether to include all in the whole shebang or create modules that allow an application developer to include what is necessary for their own application.

JIRA

The JIRA site is up and running but it’s not clear whether the mentors are waiting for the Adobe JIRA issues to be migrated before opening up issue creation to the public.

I only say this because you can get to the create issue page but no issues have been logged.

Conclusion

These summaries are meant to peak your interest if you so chose. The next step would be hitting the mail archives and getting the full deal by reading the entire thread and sometimes surrounding threads.

Mark Mail – flex-dev

Thanks for reading,
Mike