June 2005 Entries
Scott Guthrie sheds some light on the Atlas Framework that will allow for easier creation of XMLHTTP enabled applications.
Hopefully this will make the difficulties associated with Behaviors, Ajax.NET and all the other techniques as simple as
did for making Weservices easy to use.
As we move closer to Indigo, this all might be just a transitional methodology, but one that will allow us to do things for current production applications that we have a desperate need for in a disconnected internet world. It also is a must if you need to render portions of complex pages without relying on caching and postbacks.
About a Year and a Half ago, I posted this comment during the great StoredProc vs. Dynamic SQL debate between Rob Howard and Frans Bouma. I was at the time in favor of Stored Procs.
After much work, frustration, observation and rational thinking, I have completely reversed my opinion.
The biggest reasons I have reversed that opinion is because I am moving toward embracing a Service Oriented Architecture and Connected Systems. This doesn't have a lot to do with the whole debate until I started generating my Business Logic Layer and Data Access Layer with LLBLGen Pro.
Granted I was sceptical, I strongly debated whether to use CSLA and Stored Procs or a DQE (dynamic query engine). In the final analysis, I chose the DQE.
Why, you ask? The reason is simple, doing this:
WHERE
(@MyParam1 is NULL or MyParam1_Field Like @MyParam1) AND
(@MyParam2 is NULL or MyParam2_Field Like @MyParam2)
or this:
WHERE
MyParam1_Field = COALESCE(@MyParam1,MyParam1_Field) AND
MyParam2_Field = COALESCE(@MyParam2,MyParam2_Field)
is just stupid, especially when you have a dozen parameters against a table with a million rows.
It's ugly programming any way you look at it and it is just a hack.
Not only does the performance suck, it's hard to maintain (even when generated). Furthermore, in CRUD StoredProcs you are still using a procedure that is saving a whole record instead of just the fields that have changed.
When I need to save a complex data object with elements from different tables, making a Stored Proc do this efficiently and passing back validation/locking errors, etc. is something my DAL should be doing and not something actually inside the database.
When I started looking at my database as just a persistant storage mechanism and not part of my program, then ripping out all the business logic from inside the database just makes sense.
If you want to connect to multiple Database backends, then this starts becoming an even bigger priority. Sure a few of the other players support StoredProcs, you are still going to have to generate them all for each database vendor.
Why would I want to go through that pain when I can just let a DQE do it for me and not worry about maintaining change scripts.
Security is another big issue people throw up when talking about SPs vs. DSQL, well, for the last couple of years, I have ONLY used Integrated Security in my connection string so that is no longer a selling point for SPs any more either.
So I admit it, I have flip-flopped. As we move towards an even more disconnected and distributed world the logical layers we use to architect our applications are going to end up in the hands of automation tools and hopefully hand-tooling things like Stored Procedures will become a piece of architectural history.
Thursday, June 23, 2005 I will be speaking on the following subject at a special AZGroups Meeting.
---
O/R Mapping with VS.NET
O/R Mapping, or Object Relational Mapping, is the general term for the concept of creating mappings between tables and fields with their Object Oriented Class counterparts. Doing Business Logic and Data Access Layer coding has become extremely complex, especially for the developer who needs to write and maintain both the database code and the presentation code.
In this presentation I will demonstrate, by using a tool called LLBLGen Pro, how you can dramatically reduce the time it takes to create these layers in your project without writing freehand code. In other words the code is generated for you. Code Generation is just part of the O/R Mapping process, The process also may change the way you look at producing Stored Procedures and seperating your Business Logic completely from your Data Tier (SQL Server).
When utilizing an O/R Mapper in your application you are just using Objects. You will learn how all the templates work, how to create your own and how to access the Database in completely different ways than you may have done in the past. These methods are completely secure and portable from one database backend to another as well as from one project to another. The Demonstration Project will show you how in a typical database application you can generate over 200,000 lines of code.
Do you use Courier or Lucida Console for reading your code?
I find them both extremely tiring on my eyes. Roy Osherove has a pretty nice comparison of some of the different Coding Fonts available, actually any Monospaced font is fine for reading code, but most of these were specifically designed with reading code in mind.
Instead of just pointing you to Roy's post, I thought I would clarify that I DO NOT use these two fonts any longer. I am using MS-Studio's Anonymous and Bitstream Vera Sans Mono for reading monospaced text. These both happen to be on the list, but the fact that they are free was not mentioned. I have found them both to be much more clarifying with slashed 0s and a more pleasing shape the the hard blockiness of the standard fare.
Also, if you have a CRT, turn Cleartext OFF, it makes everything fuzzy without an LCD.
I couldn't stand it any longer. I broke down and bought a new phone. My Moto v60 has served me well for almost 4 years and I just didn't see any need to replace it.
Now that I am out and about a little more, I have suffered from the lack of email, wanted to go browse a webpage when I had no access, and all the other typical things someone used to being constantly connected to the internet wants to do.
I think one of my driving motivations was that I have moved all my code that touches a server into either webservices or WSE. That being said, It means that I can develop multiple clients at any time for web, smart client, etc. and have access to my data anywhere in the world, provided I have a phone that makes it possible.
I considered getting a wifi account since I almost always have my TabletPC with me, but there are still times when I don't want to have to go looking around for a Starbucks, etc.
Recently my provider lit up my city with a new EV-DO network and it now covers a pretty extensive portion of the US. This made my decision final, I can now have acceptible Broadband access with the provider of my choice, all I needed was the phone.
So Saturday, I marched over to the store and bought one. The Audiovox XV 6600. This phone is really nice, it not only has great reception, it is also a PocketPC which I have also avoided like the plague since I bought a TabletPC, I really didn't see a need for one. Now that it is built into my phone and my Tablet can take to the phone it started to make some sense.
So far I have nothing but praise for both the Phone and my provider as far as what the phone actually does and how it is working. Now I suppose I am going to have to start writing some .Net Compact Framwork application to show off the goodness and salvation of technology that is available today.
Julie Lerman led a great Bleeding Edge Session
Some good topics that came up were...
How to you handle managing time when working with so much new material?
I replied that writing EVERYthing down weblinks, conversations with other developers, etc. and using OneNote helps tremendously when doing research and coupling that with Outlook for Time Management has been essential as my schedule started to become out of control.
How can you justify working with technology that is a Year or more away?
The concensus here was to find a Business reason to move in that direction. For some it wa to develop a new product for a future platform, for me it was to take a project that is a year or more away from completion and anticipate the benefits of righting it with Beta code in the beginning and transitioning to RTM along the process because we will then have a more timely product that will have the benefits of the new technology as well as reduction in time spent developing because the new productivity enhancements will more than make up for the time spent learning the differences.
It further extends the lifetime of any product you are currently developing if you are using either a new technology that you are sure will be beneficial to the success of the project and/or think about how maintaining the product will be remarkably easier if the technology used to build it is current and you are not forced to swtch back to a non-current product to maintain that system.
It would have been nice to see a few more people in that session, but those who were there offered great talking points.
One decision I was determined to make this year at TechED was whether or not I would transition the bulk of my development to C# when moving to VS2005.
I can now say EMPHATICALLY that I am Sticking with VB.
I based my decision on 5 Key language features:
- Generics
- Using Statement
- Refactoring
- Operator Overloading
- XML Documentation
These were the ONLY things that C# had that attracted me. When you get down to the IL, the compilation of the two languages are essentially the same. I have talked to many other long time VB Developers and in addition to the above they also offered a few factors such as C# is essential a version 1.0 language, and VB is at Version 7 (version 1.0 has some trouble with things like memory leaks, that have been documented but not really made public).
VS2005 gives us C# 2.0 and VB 8.0 so, as we all know and joke about, Microsoft never seems to get anything right until version 3.0, so I'm passing on C# until then. While I may dabble in it just for fun, I won't be doing any serious development with it.
Another issue I have with C Style Languages is readability, if I put the same function written in VB and C# side by side, unless I am commenting the daylights out of C#, the VB Code is VASTLY easier to read and see were all the matching closing statements are.
Since I have been reading VB code since Version 1.0, it is almost like reading English to me; so with those productivity advantages, it will remain my Primary Coding language through the transition to VS2005 and possibly even the next generation unless there is a clear advantage to using something else. All I see now are Disadvantages for my productivity if I moved to anything else.
The rumors about Customers demanding C# is pure propoganda. In fact we recently accepted a development project in which the Customer demanded VB when given the option of 3 languages for the project. Don't believe the rumors that clients will look down on you for using VB.NET as your language of choice. As a CTO, given all other things equal, I'd much rather have a developer on my staff with 8 years of experience in VB than only the 3 possible with C#.
Doug Purdy gave an excellent seesion on Serialization.
His 3 points for cross type Serialization were:
- XSD to CLR is difficult
- CLR to XSD is straightforward
- CLR to CLR is really easy
If you are doing CLR to CLR then you probably use the BinaryFormatter. Doug was pretty specific that this is a good idea as it is the fastest of the 3 options you have for actually doing serialization now. In addition to this, he stated that Indigo's serialization is built on the BinaryFormatter.
XMLAttributeOverrides was covered by showing how to do transformations of things like what to Serialize using an underdocumented feature called ShouldSerializePropertyName(), where PropertyName is the name of the Property or Field you want to suppress sefrialization on, this is essentially equivalent to using the directive in your class, but it would allow you for example to put all the suppression in one place in you code instead of on the top of each individual property. A Link describibing the technique is here... so you just need to create a method to use it, you just implement the method and return false.
At least that is how his demo code was showing it, unless I wildly misinterpreted it. I'll have to look into this feature more because suppressing with the directive is something I have never liked and didn't know this feature existed, I would like to do all my supression in one place. You can do this with one line in C# but it may take 2 in VB, I will look into this more later.
Another really wild thing is how the XMLSerlalizer actually does its serialization... It just spits out C# code, compiles it in the background and hands you the class. This is why Serialization takes a performance hit the first time you Serialize an Object.
If you put <XmlSerialization.Compilation value = 4> in your Web.config or App.config, it will leave the code in your temp directory. If you then replace the calls to serialization with this code, you can avoid all the compilation performance hits. Very nice tip for final performance tweaking before you distribute an app. Don't forget to take that directive back out of config before you ship though.
The Swami of Connected Systems, Don Box, delivered another great session, Don seems to be getting into to being a real philosophical presenter and a lot less about down and dirty code demonstration.
I think his talks are great and provide some real direction into how to look at things when developing by associating them with the real world. All too often we overcomplicate things by trying to make sure we cover everything.
Occam was right, the simplest solution is usually the best one, especially when it comes to thinking about data structures and messaging systems. Maybe if people like Don who draw a lot of attention keep beating this drum we will get some standards that people can actually understand when they first read them.
A little tidbit of advice he proclaimed, was to write everything down... Many of us tend to do alot of things in our head, especially planning or thinking about future ideas. I have found OneNote and my tablet indispensible for this.
I had the pleasure of finally meeting DonXML (Don Demsak) as he lead this Birds of a Feather session. Another surprising attendee to see was Matt Curland (of VB's Adressof among other thing's fame).
While XSLT2 and XQuery are not yet approved W3C standards, there are still some reasons to use them. Also Discussed were the XME Extensions and a host of other useful tools available with sorcecode at http://sourceforge.net/projects/mvp-xml
I was fortunate enough to receive a free XSLT2 Book written by Michael Kay, will be an outstanding reference... His XSLT 1 book was indispensible to me as well.
Another favorite presenter of mine is Clemens Vaster and once again, this year he did not disappoint.
Asynchronous stuff is really hard to program (correctly) and he showed some nice misconceptions about how some things that coders do and think are asynchronous, actually are not... Such as spinning up an animation while you wait for a page to load is definately not asynchronous, it's two synchronous events happening simultaneously
Asynchronous means you aren't waiting around for a response at all, you just do something interesting when a response happens to be received.
With Reliable messaging there are a couple of ways to handle state, Volitile RM and Durable RM, with Durable RM everything gets written to disk (or database) during each step of the message's transition, this will survive network interruption and hardware failures and allow the message to restart it's path where it left off.
One way of doing this is through the use of MSMQ, which also happens to be the one I use, Programming this technique with Reliably and handling errors gracefully on both client and server is really, really hard to do correctly.
I have goine through the pain of writing all these controls for a proprietary system, but Clemens is good enough to issue his tools with a BSD License for using MSMQ to do the Durable Messaging and are available at his blog.
One thing he said a few times that is getting to be a frustrating worry of some developers I haved talked to is the complexity that XML has now become. I agreee, when XML was relatively new it was cool and fun to go muck around in it, but when it comes to useing tools like SOAP, forget it, you are insane to writethose things by hand...
Trust you Object Models to do all your serialization for you and let the other tools that can generate XML do it and stop trying to understand the plumbing inside the XML unless it is something you really need (like you are the plumber...)
Just because XML is suppossed to be Human Readable Text doesn't mean what it contains is something you need to go understand at every element and attribute everytime you use XML as a transport mechanism.
A few things gathered from the WSE3 What's new session...
MTOM - Message Transmission Optimization Mechanism
- Replaces Dime and WS-Attachments
- Secures Attachments as well as message
- composes WS-Security
- wire level reduction in size
- Keeps the conversation alive So you are not, in effect, redialing the telephone for each question and response
- even when state and wire are volatile the state can remain open for a given time (Reliable Messaging)
- Stateful SCT for stateful conversations in a web farm WSE2 cannot to this easily
WSE3 is communication compatible with Indigo
WSE2 & 3 can run side by side
Migration tool for WSE2 to WSE3
Remember that WSE2 cannot interop with WSE3 or Indigo
I think this is the 3rd time i have seen Michele Leroux Bustamante do this presentation and while she is a delight to watch and listen to as a presenter... WSE2 still sux.
Yes, having a stadard to work with is great, but this stuff is really long winded, the wizards don't cover enough to be truly operational and WSE2 is NOT going to interop with WSE3 or Indigo...
So What? I have to use it anyway, but I sure don't like it, doing raw SOAP calls is a heck of a lot worse, but you can do many things just using the asmx API if you don't need to gurantee standards conformance to a client.
I really look forward to going to WSE3 soon and doing more with WS-Reliable Messaging
Michele covered pretty much everything you need to know to be really dangerous in WSE Security and showed how to get certificate negotiation going, this stuff just can't be this hard to implement if it is going to get widespread adoption.
ADFS - Active Directory Federation Services
This was formerly known as Trustbridge and was announced late last year. It is coming in Windows Server 2003 R2, an overview is available here
It is supposed to get us further towards a Single Signon implementation and from what I have seem looks pretty good in the way it implements Claims to SIDs and vice versa.
.
MIIS is the tool you can use now with ADAM (active directory application mode) to get some of this.
ADFS GREATLY simplifies the current hassles of integrating ADAM and MIIS into your applications. ADAM is something you should definately be looking at if you need to do AD Authentication, I am currently doing a lot of this through LDAP and it's not very easy to implement, this may take some of the pain out of it and it was stated that the implementation is the same even in Longhorn so if you understand how ADAM works, you are learning the right methods today.
See: www.identityblog.com for how Microsoft (Kim Cameron specifically) sees the 7 Laws of Identity
#1 is User control, thus the reason people HATE, yes HATE RFID.
Microsoft seems to think there will arise a Meta System for Identity (which makes a lot of sense) similar to the way TCP/IP is the Meta System for networking.
Indigo, InfoCard and Active Directory will all support this Meta System when it arises (supposedly...) but i will believe that when I see it, I definately hope that this shows up soon because i am getting really tired of implementing 3 or 4 different techniques in just about every program I write.
In VS2005 there are some Significant Changes coming to the System.Net namespace.
These changes include:
- HTTPWebRequest
- FileWebRequest
- FTPWebRequest
- SMTPClient
SMTPClient Replaces System.Web.Mail and is the most significant change to the namespace.
SMTPClient supports Attachments and Views, Embeds Authentication and SSL, can be used with the IIS Drop Folder and provides MUCH better scalability.
A few people have commented to me that these new abilities make thrid party tools like AspNetMail less attractive because most of the same abilities are now "in the box". If you need fine grained control of an SMTP service, you are still going to want the third party controls, but for most activities, like simple mailing lists, logs, reports and general mail, this is going to be more than sufficient.
The FTPWebRequest now adds much better control over sending and recieving via FTP even FTP Over SSL, it opens up most of the options that are needed to even build an FTP server and will certainly enable you to build a decent client now. These Objects no longer depend on WinInet, they are implemented at the TCP/IP socket level. You can also confirm at the process level now to support multithreading, SSLStream can negotiate authentication based on certificates and NegotiateStream can talk to Active Directory.
These are great additions to the tools we are getting in VS2005 and I look forward to utilizing them soon.
In the Day 2 Keynote, the only new Item of note was the Oficial Release Date of VS2005 and SQL2005: November 7, 2005
So now you know when D-Day is for the official release
This wasn't part of the keynote, but it was also said "officially" that the release of WSE3 will be 2-4 weeks AFTERthis date, but may be sooner. Note that WSE3 is not forward compatible with WSE2 or Indigo, I will cover that elsewhere.
If you want a real blow-by-blow, read Peter Galli's fine coverage...
Some thing I found really strange was the announcement of Per Application Firewalls. This just sounds like a dumb idea to me, sure I can see some possible uses for it, but I think it's more like a nightmare in the waiting.
A footnote that does look like really interesting technology is RDC (remote differential compression) Basically this is a way to transfer the changed part of a file instead of the entire file, sort of like a patch file for any type of file, cool stuff here.
This is a new fairly new announcement.
It is New program to differentiate the "Cream of the Crop" of Architects.
- Pay to get in Program (a lot... several thousand dollars...)
- You will get a mentor to guide and assist in your training
- Must agree to be a mentor after beign Certified
- Board to oversee Certification
Two types: Solution Architect and Infrastructure Architect
Disparate systems and Interop Technical know-how are encouraged and in some cases required.
Program goes Public in January of 2006
This program looks really good, it will be along the lines of Regional Directors and MVP, but much, much more. Since you are reviewed by a Board, there is just no way of cramming for a couple tests. It will give a lot of validity to the program and there will be some great opportunities to come out of it.
Mr. Ballmer is in "Prime Form" showing off the latest and greatest from Redmond
"The New World of Work"
hmmm, ok... Who comes up with these buzzlines?
Find the information and communicate it.
Yes, this is pretty much what we have been trying to do with computers for the last 30 years...
Sometimes it really seems like these are just responses to Google and other competition, I'm sure some of that is true, but you know that Microsoft has been working on this stuff at least as long as some of the others, they just happen to be coming out around the same time. It's not like you can reverse engineer Google maps in a couple weeks even if you are Microsoft.
Samantha Bee does news bits, this was moderately entertaining
Top 5 Requests by info workers
- One Identity
- Online Presense
- Network access
- Synchronize
- Self Service Websites
- + Rights Management
Sure I can agree with these, and the following offered solutions are pretty good, but certainly not the only ones to consider, even today...
- SMS
- Sharepoint
- Exchange
- Active Directory
Yep, they can make this happen, but when Exchange 2003 chokes on a single attack from spammers and LSAS continually shuts down from overload even in a small set of 3 domains (as recently happened to one of my systems) then I think Exchange still needs either A) alot of work, or B) too many 3rd party products to make the solution optinmal today.
Sharepoint is great, and it's something I really like. I use WSS all the time, but I would like to see it be just as usefull with an easy way to store the shared files outside of the database as easily as it is to store everything inside the database.
With that option, using Sharepoint to share millions of files could be really fast, when you have that many files in anything but Data Center edition, it's just too slow. There really isn't any need for that, If I could just add new shared storage as the needs grew it would be a better implementation for smaller operations that know they will grow exponentially, but need to start with smaller hardware to ramp up to the larger stuff as revenue comes in.
W2K3 R2
- AD Federation
- Trust between domains
- made simple to implement
- Active Directory Single Sign on
- WS-* extensions
- Virtualization will be Builtin
- Clustering Server with distributed task scheduler
Excellent Additions to the most stable and reliable Windows Server Yet.
Exchange
Direct mobile Messaging
Exchange 2K3SP2
Direct Push Email
Can wipe remote clients
This was a nice demo, but I still don't trust Exchange, and I sure wouldn't trust Everything to it by adding more before there is a better way to control Exchange. I have a MUCH more reliable Linux/Postfix/Courier/Spamassassin/MySQL right now and will continue to use it while I see Exchange blowing up at normal network nightmares like heavy spam attacks. However, making Exchange an internally relayed server that provides all the great uses of Exchange, without actually being open to mail traffic works pretty well. (Yes, indeed, strange thing to say about a supposed Enterprise Class Mail Server) but that's just my opinion and I am not an Exchange Guru, I am sure some people can throw enough money and time at it to make it work easily, I can't.
Virtual Earth
Just Like Google maps but better... (Mr. Ballmer didn't say this I did...) It was a really cool demo nonetheless.
http://virtualearth.com and it seems to be down right now (or not yet available)... Here is an earlier take on it... http://blog.searchenginewatch.com/blog/050523-125208
VS2K5 for Office
With VS2K5 for Office, Outlook can act as a Command Center for your apps and give you time critical disconnected apps hosted right inside Outlook, this is a great Idea that started with Dashboard, then became Digital Dashboard, and has finally become useful.
MOM can now manage Sun as easy as Windows (at least they made it appear that way in the demo) Very cool for us interop folks who want that sort of thing.
Windows is now more secure than Linux, at least some people really believe that, it's not just marketing hype any more.
New MS update
More than just an update agent roll out patches by schedule or SMS
This is pretty nice, I look forward to using it to control scheduled updates easierin a small office, SMS is usually overkill in those situations, but it can be applied now in smaller (non Enterprise) environments with just AD Policies.
Overall a very inspiring Keynote... They also showed an OOOH and AHHHH demo of the search capabilities of Longhorn in a way that looked so similar to Photo Triage you have to wonder if it is just smoke and mirrors right now since it is still at least a year away, surely it was nice and supposedly going to be extensible. Quite inspiring, but a we have to wait for it kind of thing.
As I prepare for TechED, I was wondering how I could best spend my time this year...
I spent a lot of my time in the Cabanas last year talking with really smart folks and I will probably do the same this year. I met some really great people last year and hopefully I will meet some more this year.
One thing I definately will be doing different is spending a bit more time in the C# sessions...
If you are wondering... no, I am not abandoning VB development, I just supplement it occassionally and since I have moved more into architecture than development, I need to "broaden my horizons" more.
For me, I think the hands-on-labs are a complete waste of time, now that we have VPC I can do all those on my own.
Last year the Architecture track was so high-level it was a 50,000 foot view and useless for anything but general interest. So I will be sticking to the developer track and hopefully get some meat as opposed to overviews of "what's new" in VS2005.
Birds Of a Feather Sessions of interest: Software Factories, Asynchronous Request Processing, Test Driven Development, Avalon/XAML, Preparing for Indigo, and Life on the Bleeding Edge - notably missing was a Tablet PC BOF that was great last year.
Only 36 Sessions total for the Tablet this year... hmmm, why is this not being promoted more heavily? Most of these are even just Mobile computing in general and not actually Tablet specific. My take on this is: Supoort the Developers or no one is going to develop for that platform... come on MS, get a clue, you can't say how great the Tablet PC is and how ALL notebooks will be tablets if you aren't going to promote development for them more. I'd personally love to have a Desktop that is a Tablet... a 30" Apple Cinema display with a Digitizer Interface would just be so sweet.
Blogroll Me!
Blog Search Engine
Copyright © 2003-2004 H. Steele Price, IV -
All opinions are my own, not necessarily those of my employer, your mother, or any government agency.