The paper argues that functional programming provides two important features that incredibly useful:

• High-order functions
• Lazy execution

I agree all the way, and having experienced and heavily influenced myself with the way functional programming strucutred and modulirazed codes, I do agree that people need to look into functional programming and realize how it can be very helpful to them.

Why Functional Programming Matters

Thank you for using my Tabelog Client app for Windows Phone 7. I hope you’re enjoying it!

(English follows)

検索画面

検索画面で「キーワード」と「エリア・駅」入力し、「検索」ボタンを押すと検索実行になります。

キーワードにはレストランの名前（一部含む）、グルメの種類など入力可能です。「エリア・駅」にはエリア、都道府県、駅の名前などの入力可能です。

気に入り

発見した店を保存する機能は入っております。いつもくる店、行く予定店、これから計画して行きたい店を気に入りリストに入れると簡単に見れることになります。一覧に入ってる店をタップするとすぐに店の詳細を表示されます。便利なリストです。

コンテクスメニュー

気に入りリストに表示されてる店はコンテクスメニューが付いております。アイテムを長押しするとこのメニューが出てきます。メニューの中に’pin to start’メニューはこの店をスタート画面に貼り付けという意味です。ピンされた店はアプリ起動しなくてもスタート画面から店の名前、営業時間と電話番号が表示されています。とても便利なツールです。ピンされた店をタップすると店の詳細が表示されます。
もう一つのメニューアイテム’remove’を押すと選択されているアイテムを気に入りリストから削除される。

履歴

履歴リストに検索されたキーワードとエリア・駅をリストアップされます。この画面で同じ条件で検索するのは簡単にできます。

検索結果

検索結果はリスト上にまとめてあります。

エリア絞り

エリア特定し、さらにエリアで絞りたいであればエリアボックスをタップし表示されたリストの中から検索したいエリアを選んだらそのエリアに該当な店が表示されます。

コンテクスメニュー

検索結果のアイテムにもコンテクスメニューがついております。このメニューを利用して、気に入りリストの追加とスタート画面に貼り付けするのは簡単にできます。

詳細

店の住所、営業時間、休日、電話番号、値段などの情報は詳細画面から調べることができます。検索結果リストからアイテムをタップし、この画面が表示されます。同じく、スタート画面にピンされた店をタップするとこの画面が表示されます。

メニュー

詳細画面はコンテクスメニューではなくアプリケーションバーメニューで機能が表示されます。

スタート画面に貼り付けと気に入りリストに追加機能以外に、ブラウザーで表示するメニューがあります。このメニューを選択するとInternetExplorerでこの店の食べログページが表示されます。

口コミ

店の口コミはこのページに表示されます。

食べ物/店内/店外/その他

店に関する写真が表示されます。

———————————-

Search Screen

The search screen is where you specify the search criterias for venues to search. It has two primary items to consider, which are the ‘Keywords’ and ‘Location’, although both of them are optional.

A keyword can be a complete or partial venue name, or type of food, or any other word that may describe a venue. While location are place in general where to locate the item. Location may be an area name, prefecture, even a station name.,

Favorites

Occasionally you will find yourself browsing through search results and wish you can store some of the venues for viewing later. Or, you want to keep a lits of restaurants that you think is fabulous to visit in the future. This is where the favorites list come in. In this screen, you can find all your favorites venues and access their detail view.

Context Menu

The items in this screen provide context menus which provides additional functionalities. The ‘pin to start’ menu pins the selected venue to the start screen. And the ‘remove’ menu removes the item from the favorites list.

Recents

Recents keep queries you entered in the search screen. The screen provides a historical list, and by tapping on an item you can easily search using the exact same terms listed.

About

Search Result

Search results are displayed in a single list showing the venue name, general area and overall rating. Tapping on the item opens the detail view for a venue.

Area Refinement

Often you have too many results to pick from or want to refine the search into more specific areas. The area refinement box is the way to refine your searches. When you tap on the box, it will give you a list of areas to look for. Using this mechanism you can drill up and down of areas. The area refinement box shows the current area by default.

Context Menu

Venues returned by the search also have context menus attached to them. User can easily add items to favorites and pin the venue to their start screen.

Detail

Detail screen is where informations from a particular venue gets displayed. It shows address, phone numbers, working hours, holidays, prices and other informations. Tapping a pinned tile on the start screen immedietly opens this screen for the item pinned.

Menu

The detail screen uses application bar menu to show options instead of context menus.

Other than the usual ‘pin to start’ and ‘add to favorites’ menus, a new menu ‘open in browser’ is added. Selecting this menu opens the tabelog page for this particular venue.

Review

The review screen takes reviews for the particular venue and show them in a long list form

Food/Inside/Outside/Other

These pages shows photos for the particular venue.

bjam --toolset=msvc --build-type=complete --with-regex stage

That solved the problem. At home, and at the office.

http://en.wikipedia.org/wiki/Security_through_obscurity

I wonder what kind of reaction will come from other browser vendors? Biased because sponsored by Microsoft? Too cheesy,…

Here’s the link to the Microsoft Office Labs Envisioning Videos:

http://www.officelabs.com/Pages/Envisioning.aspx

Here’s a link to a blog that post the full video of ‘a glimpse ahead:

http://www.istartedsomething.com/20090228/microsoft-office-labs-vision-2019-video/

UPDATED: Here’s the full video in youtube:

http://www.youtube.com/watch?v=XiqgmAYrd3c

As I watched the video and awed on what is possible in the future, I can’t help imagining on the technologies that is being developed right now to enable those solutions. First is the bendable screens. These screens able to display graphics and movies on a piece of flexible plastic with superb colors. You can also roll these screens to somekind of a parchment that will absolutely saves spaces. Another fantastic property of these screens is that it also can be an input device, or in other words it can also act as a touch screen. Now, think about if our computer screens, televisions, cell phones, mp3 players, or even tables and walls are using this technology, now suddenly we have access to information everywhere at home. The technology itself for bendable screens are already been invented and right now we are only waiting on refinements of the technology and for companies to capitalize on it.

Next, is internet connectivity anywhere, anytime. Demands on being to stay connected anywhere and anytime has grown bigger and bigger with every year. I still remember fifteen years ago, I don’t have a cell phone and does not feel the need for such devices. Fifteen years ago also, I did not know anything about the internet and was playing around with Turbo Pascal on my 486 system. But look at where we at right now, cell phones and the internet are technologies that can’t be separated from our lives anymore. Here in Japan, everyone has a cell phone, even trends here start for people to have second cell phones and so on. Not only that, here in Japan we have internet access directly from our cell phones. Latest generations of cell phones includes full internet browser capabilities and such.  Not only cell phones, Wi-Fi spots have been a trend lately and cafes, airports starts on adding wifi points to their service. Wifi connections also making their debuts in Schools and Universities. Looking at the trend of internet, we can be certain that making people stay connected anywhere and anytime is an ongoing effort, but it may require a bit of time to reach people that is outside of facilities. Nevertheless the future of stay connected people is coming closer and closer with each day.

Next, with constant internet connectivity, the need and value of having large disk spaces and superior computing power becomes questionable. Lately, cloud based storage and cloud applications are being promoted by big vendors such as Amazon, Microsoft and Google. Their solutions for cloud applications enables users use their applications and access their data anywhere as long as Internet exists. Technology such as Microsoft Azure platform delivers the framework on enabling such solutions. The main point here is that the value of ownership of a storage device is diminishing, the value for owning high performance cpus are also diminishing. What is left is probably UIs. But with data and applications being stored in the internet, having UI devices might not also be as valuable as owning a 1000 dpi laser mouse right now. Another interesting point about these UI devices is that, if you relate to the first hardware advancement that I discuss on the top is that these devices might only be a piece of displayable plastic with some kind of controller and internet receiver (and probably a camera and a mic). Those can be very cheap to produce and I would not be even surprised if people can exchange and reuse the same devices.

Next, Augmented Reality. The ability for computer to recognize objects and add informations on the object will opens a lot of way on how we collect and use informations right now. Gone are the days when you want to know the name and nutritions of a fish and spends the night surfing the web at home based on a picture you took at the fishmarket. With augmented reality, you just point your computer camera at the fish and boom the computer will recognize the fish and shows informations about the fish.

Next, OpenId. This is a vital technology to enable cloud computing and collaborations to work seamlessly (other than web services of course). The technology enable  a person/entity to be recognized universally in the internet. Integrations between services and networks will rely very much on this technology. Having an open ID is like having your Driver’s license or, in my country, your Citizenship Id card. Your Id will be accepted nation-wide and people recognizes that Id as proof of what your claims is. OpenId provides the framework to do this in the internet. Now, with OpenId you can stop inputting different username and password for different sites. In fact, you may not need to input anything other than showing your OpenId to visit a site anymore.

There you go, a list of  tehcnologies that I think will drive what Microsoft envision today. Of course this is not a complete list by far, however these are the technologies that I’m interested in the most.

Control Execution Lifecycle

ASP.Net Page Lifecycle

As usual, I always drew designs before I started programming. This process actually took me around 3-4 hours. And in this design phase I quickly identified problems. Given my experience with ASP.Net, I know that controls automatically can handle postbacks and set their values or state to a correct value. That is, if the controls are static. But what about if the controls are dynamic? Well, as long as you reproduce exactly the controls then it wont be a problem. But what if you cant? The problem with questionaires is that you may have lots of questions that can span to multiple pages. AND, because this is general control, it means no statically defined pages or control tree (unlike the cross page posting where you can refer to the previous page in strongly typed manner). Some browsing in the internet for this problem reveals a solution, it is to rebuild the control tree and then redirect. Another solution is to use java script which will call a special method that will update your data. None of this is what I really wanted to do. For the first one, it seems terribly stupid and I have doubts that you can cleanup your controls after the Page Load event is over. It looks like a hack and I wanted to find another way for doing things. THe second one, means I have to build a special service for that and that is high coupling and have to be done outside of the control border. Another reason is that it is completely relies on client browser doing the right thing. I want to avoid browser compatibility issues as possible.

So thats the dilemma, after a while of thinking, I thought on an idea of actually dealing with posted data manually. If we can make the key value of the posted data predictable, actually we can just tapped in the postback data and mapped them to the entities. The key for this scheme is the Id. This Id have to be unique and can be mapped back to the entity in question. An important note to remember is that even though I said it should be able to map back to the entity, but I realized that is not a good idea from security point of view. What if a hacker manually created his posted data and modify the id to modify some other question? So here is my scheme for this Id:

Client            |          QControl          |   Entity
------ Id ---->   |  Reconcile  <-- Predict Id | <-- Id ----

With that out of the way, my plan is to build a simple questionaire first that only consist of Labels and Textboxes, and all the layers that is needed to support it (means building entities that the web questionaire control can understand). Very simple, very straight-forward. The web control to request the list of these predefined entities and build controls that will reflect them.

public class entity {
public String Id;
public String Question;
public String Answer;
}


Now, the questionaire control can automatically build the controls that will represent those entities. The first question is when to build the controls. ASP.Net Page triggers events to indicate the position in the lifecycle where they are at. Now, we can trap these events and build our controls right there. However, then our control will be tied to the Page and that will make the control unusable. What we need is something more fundamental.

As I mentioned before, I am trying to build a control, and as all controls in ASP.Net is derived from Controls (or its derivatives of WebControl or CompositeControl). I started by deriving my control from WebControl.


public class WebQuestionaire : WebControl


From there, I went forward on creating dummy data so that the control available in the designer preview window and so that I dont have to worry about the data at this stage. The implementation of the control building was pretty straight forward.


protected override void CreateChildControls ();


This method actually handles the control building that ASP.Net will call on creation of the control. Why not build this in the Questionaire control? Some of the things that I noticed if you do that in the constructor, then the base control tree within the page itself was not ready yet. Other than that, properties in the control itself was not been set and will have the value of Null. THe Page property have null because the control is not being created yet. So, instead of having to handle that pain, just do something simple to initialize the control itself and let ASP.Net request for child control creation when the time is right. Now, why do I need access to Page control. It is to get to the main Form property where the postback data will be available. Another advantage with waiting the creation of the child controls is that I can get the definite control id for building my unique ids.

In order for building a recognizable id, I have to account for duplicates. This can be handled by having the Questionaire control be marked by INamingContainer:


public class WebQuestionaire : WebControl, INamingContainer


All child controls of this control now will have the control id prefixed on the front which will eliminate the duplicate problem. Also, we can get this id programatically by using the UniqueId property or the ClientId property (for their differences refer to MSDN).

Now, I can create the controls, I can create the unique Ids, I know that this is best being created using the CreateChildControl method. BUt what about the Id reconciliation what is the best place for placing this? I am not sure if this is the best way, but right now what I did is I hijacked the events. Controls actually expose some event trigger methods that can be overriden such as OnInit, OnLoad, etc. What I did, I override the OnLoad (maybe I can do it using OnInit also, but I fear the controls not being initialized properly if I mess around with it. But let me try that another time) and access the Page.Form to access the postback-ed data and run my reconciliation method there. And it worked fine. So now, I have a working protoype on the Web Questionaire Control.

I realize there are some points that need to be researched more before I can go further. So, right now Im going to reanalyze things and weighs my options.