Vote of No Confidence

EF Vote of No Confidence

Elisa Flasko's Blog — Tue, 24 Jun 2008 19:59:40 GMT

There has recently been an interesting thread in the community going on about a "vote of no confidence"

re: Vote of No Confidence

James Knowles Blog — Wed, 25 Jun 2008 01:12:56 GMT

Am I missing something is a vote of no confidence the way forward...

re: Vote of No Confidence

joewood — Wed, 25 Jun 2008 03:41:21 GMT

Interesting post. I think it would help to provide some timescales around v. next - even in CTP form.

I'm also interested in why the model file needs to be anything more than the class model diagram in Visual Studio now. This would at least address the code merge issues.

Thanks for the post

Jeff G — Wed, 25 Jun 2008 05:05:14 GMT

Thanks for the reply to the "no confidence" article. This has cleared up a lot. Support for Automatic Lazy Loading is what brought me here. I understand why you did it that way for v1, but just makes for too much code. I'm looking forward to seeing what you do here as well as on the domain model story.

ADO.NET Entity Framework Vote of No Confidence

BlogCoward — Wed, 25 Jun 2008 05:17:55 GMT

ADO.NET Entity Framework Vote of No Confidence

Entity Framework vote of no confidence

Wed, 25 Jun 2008 06:21:41 GMT

As was announced initially (as far as I can tell) on Bil Simser's blog, some concerned citizens of

Couple Responses to Comments...

timmall — Wed, 25 Jun 2008 07:41:53 GMT

>>> I'm also interested in why the model file needs to be anything more than the class model diagram in Visual Studio now

When we support POCO I think this is much more doable. Today the EDMX file encapsulates a lot of artifacts that we can infer by convention or via attributes int he POCO scenario.

>>> Lazy Loading

We definately need to work on this in V2 we will throw thoghts up on the EF design blog

Tim M

re: Vote of No Confidence

Wayne — Wed, 25 Jun 2008 08:01:24 GMT

blah blah blah, most developers are still working in 1.1 or 2.0. WTF is EF anyway? We don't have time to continually re-learn .net. Get it done and finalized, only then will I dig into this crap. Till then, I'm staying on 2.0.

re: Vote of No Confidence

Bryan Reynolds — Wed, 25 Jun 2008 08:01:35 GMT

Microsoft will get it right at version 3.

Entity Framework: Vote of no confidence... e le risposte

di .NET e di altre Amenit — Wed, 25 Jun 2008 10:42:19 GMT

Entity Framework: Vote of no confidence... e le risposte

re: Vote of No Confidence

colinjack — Wed, 25 Jun 2008 17:07:33 GMT

"What we do believe, however, is that it is desirable to have a single meta-model (EDM if you will) with which you can describe many domain models and that by having a single grammar we can provide a set of common services on any given domain model."

I'd be interested in reading more about how you see this working in practice, particularly where there are lots of applications and seperate concerns (including reporting) taken into account.

This could clear a lot of things up, for example if you have multiple domain representations of a Customer what does the association entity in the EDM look like?

Evento de EF, novedades para la siguiente version ( no os enfadeis ) y alguna cosa interesante

O bruxo mobile — Wed, 25 Jun 2008 21:48:44 GMT

Bueno, lo primero es lo primero.... Tengo que decir que he quedado encantado con el evento organizado

re: Vote of No Confidence

Brian Grant — Wed, 25 Jun 2008 22:30:59 GMT

This is yet another case of Microsoft ignoring their customers’ requests and trying to force some half balked (we will get it right in version 3) technology down the developer communities trout. In case you brain surgeons haven’t figured it out by the poor response WPF has gotten the community is sick to death of this nonsense.

re: Vote of No Confidence

lowendahl — Wed, 25 Jun 2008 23:44:53 GMT

MS will get it right with V3 and by V4 they will have the market. Just like with any other of their products :) Jokes aside, I really like this new transparent side of Microsoft. Finally their products will be driven by customer feedback from the community at large and not only by a handful of "friends of theirs" and big partners/customers. Well they are important, but there is more developers out there :)

Thanks for the reply on the vote as well. I'm looking forward of seeing the interim builds on EF v2.

Jon's News Wrapup - June 25, 2008 Edition

Jon Galloway — Thu, 26 Jun 2008 01:09:55 GMT

h2.entry-title { font-size: 1.1em; clear: left; } ul.hfeed { list-style-type: none; } li.xfolkentry

re: Vote of No Confidence

danieldsmith — Thu, 26 Jun 2008 01:39:14 GMT

This "vote of no confidence" is a tad harsh and I do feel for you guys. It must be terribly disheartening for the EF team. Please don't let this discourage you though! The goals in principle are definitely a step in the right direction and it is reassuring to hear that you're taking the community's criticisms onboard with plans to address the issues in v2. In a way it's good that you're getting so much feedback – this is obviously an important area that people are very passionate about.

That said... I do have to say it is *very* frustrating to get all excited about a product, only to find that there are limitation that make real world scenarios painful to implement.

As Brian (above) mentions, WPF is a perfect example. I can't believe that WPF only recently got support in SP1 for hooking up events in the designer! It's *basic* things like this that really should have been in from day one that frustrate developers the most.

Same goes for the EF. You already know what the major pain points are in v1. So why release something that you *know* isn't done? Is it sensible to tie yourselves to the SP1 release schedule? Why not start on v2 right now and release that as an out of band release? To be honest, I'm a little surprised that so much new technology is being introduced in a service pack.

Just a suggestion, but why not release the whole of the EF on CodePlex. Do frequent releases and let the product mature, and when it's ready, then bake it into the .NET Framework. A really good example of this approach was the ASP.NET AJAX (Atlas) project. That was a resounding success and the project quickly matured and made it into the framework.

re: Vote of No Confidence

Jon Davis — Thu, 26 Jun 2008 21:31:45 GMT

I think the fundamental issue here is that the "coding in a box, then seeking feedback on preview or final releases" is a flawed development model *when*, and only when, you're dealing with generic frameworks that emphasize developer productivity and/or manageability, like NHibernate or Subversion, rather than highlighting and supporting a proprietary Microsoft product like Office SDKs.

There's no question you guys are working hard, and kudos for your work. But this is very specifically an area where only the brightest and most experienced software architects currently living in the universe CAN lead, and if the intent is not to lead, why bother?

Great list of members in the advisory committee, but now the questions become how much are you willing to rely in their advice, and to what extent are you willing to submit to their leadership?

ADO .NET Entity Framework and the Spark that brought in the Heavyweights

Murray Gordon's Blog — Thu, 26 Jun 2008 23:18:35 GMT

I came across the "ADO .NET Entity Framework Vote of No Confidence" at http://efvote.wufoo.com/forms/ado-net-entity-framework-vote-of-no-confidence/

re: Vote of No Confidence

Peter Morris — Thu, 26 Jun 2008 23:57:18 GMT

I often hear that MS don't listen, I think your response shows it does.

re: Vote of No Confidence

Fri, 27 Jun 2008 15:57:49 GMT

"but we had to deal with the tension between trying to add more features vs. trying to stay true to our initial goal which was to lay the core foundation for a multiple-release strategy for building out a broader data platform offering".

The question of course arises whether the cart should be put before the horse. If the consumers have told you what they want, and the product does not meet these needs, why flog the dead horse? Shouldn't customer needs drive or at least guide product development?

re: Vote of No Confidence

Special Speller — Fri, 27 Jun 2008 22:26:24 GMT

Brian Grant is a Half Balked Trout Stuffer

Learn how to spell pal, then you can fling the poo.

The Cathedral, the Bazaar, and Microsoft’s presence on both sides of the Vietnam War of Computer Science

בלוג היועצים של מיקרוסופט ישראל — Sun, 29 Jun 2008 15:31:21 GMT

Let's start this post with a couple of warnings / disclaimers: Be advised that this post mixes metaphors

ADO.NET Entity Framework ja veel midagi

Iga lahendus tekitab uusi probleeme ehk alati võib leida veel ühe bugi. — Tue, 01 Jul 2008 11:35:24 GMT

.Net -i arendajate kommuuni poolt on postitatud väga asjalik "Vote of no confidence" artikkel ADO.NET

re: Vote of No Confidence

Richard Dinel — Fri, 04 Jul 2008 18:32:46 GMT

i went to read the list of people and when i press for page 2, nothing happened (loading data...) is it nHibernate performance (lol) ?

re: Vote of No Confidence

Avi Shilon — Wed, 16 Jul 2008 17:41:15 GMT

When I first heard about the vonc, I was one of those that thought that it was a little too harsh and unfair. After all, the EF is a very promising piece of technology and it's just making its first step in its 1000 mile journey.

But then I got it. The team should actually be proud that the vonc exists at all. You see, when you deeply care about something, this is when you honestly try to fix what you think needs to fixed, otherwise you couldn't care less and you'd just be ignoring it. I think the whole industry has high hopes from the EF and that's why the vonc. And who better to make the loudest resonating claim than the best the .net community has, the mvps. Most of the industry trusts them to know what ticks right and what needs a little bit of tweaking. They are eye openers for people like me who are less experienced. Without them I would have thought that the EF is perfect. And if they make a such loud call, all of us should listen.

And because of this heavyweight figures, the EF team should be proud of their baby, not hurt. The best of the best do really care about the EF that they gotta be sure that it shapes up nicely and the opportunity won't be missed. So, the bottom line is, this is all constructive criticism coming from good intentions. Keep up the great work.

re: Vote of No Confidence

ekepes — Sun, 20 Jul 2008 21:03:19 GMT

None of the other things matter if I can't reliably version my source code. The merge problem is the biggest one, and while you seem to be rather flip about it, it really makes the difference between whether I can use the EF in the real world or not. That is the only reason I signed the petition. I can deal with the rest of the problems - they are differences of opinion, and will work out over time, but without reliable version control, I may as well use Access.

David Yack will be presenting on Exploring the Entity Framework for the Tulsa SQL Server Group!

David L. Walker - .NET Solution Provider — Thu, 31 Jul 2008 09:08:06 GMT

I had to be in Los Angeles a couple weeks ago now for some cross training and to meet my awesome development

David Yack will be presenting on Exploring the Entity Framework for the Tulsa SQL Server Group!

David L. Walker - .NET Solution Provider — Thu, 31 Jul 2008 09:22:45 GMT

I had to be in Los Angeles a couple weeks ago now for some cross training and to meet my awesome development

For Open Source: LAMP; For MS: ...

Hilton Giesenow's Jumbled Mind — Mon, 04 Aug 2008 18:27:47 GMT

I'm sure many of you have heard about a very common open source acronym: LAMP . If you haven't

re: Vote of No Confidence

shah — Mon, 04 Aug 2008 19:16:10 GMT

I am confused about the need for yet another language like ESQL when there is LINQ. This is all very confusing ? I was considering adding LINQ to SQL but now you saying that LINQ can not handle this because of flaw in LINQ or performance or something. Dump ESQL altogether.

re: Vote of No Confidence

dnp — Thu, 07 Aug 2008 01:10:55 GMT

I can't say I agree with some of the points made in the VOC, although I'm not saying I'm happy with the EF either. I believe there are other area's such as a very poor story around stateless interaction which are of more concern. Also M:1 association by key isn't particularly great either. I dont buy the Lazy load, I agree with MS that it should be explicit. Granted the implementation could be worked on, but the concept is right.

ADO.NET Entity Framework: Impressive! Powerful! Useless!

Reddnet Scribbles — Sat, 23 Aug 2008 09:58:27 GMT

The new Microsoft Entity Framework is the latest in a long line of very impressive, yet tragic failures in Microsoft's data access strategy...

re: Vote of No Confidence

Qiang — Fri, 29 Aug 2008 06:58:53 GMT

I don't understand why the EF doesn't support the native SQL, and some generacal and direct DML methods.

re: Vote of No Confidence

steinard — Thu, 25 Sep 2008 15:28:54 GMT

I have used Hibernate and NHibernate on many projects and worked with ORM technology since 2003. When reading some of these comments I wonder if developers in the .Net community have experience with other ORM tools before testing EF. It is very difficult to see what you are missing without knowledge about what features the other ORMs can offer.

So, if I were to lead a conversion from NHibernate to EF in one of my current projects, then this technology replacement would cause a high cost to the project in terms of loss of functionality and a lot of code needed to be written just to make EF work like NH does now. This also include changes in the architetcure, which seems like changes for the worse with no gain as far as I can see. The points below illustrates what I loose by switching to EF:

- A default fetch plan through mappings specifying what to fetch by default. This must be replaced with a lot of boiler-plate code.

- Eager loading and implicit lazy-loading, as everything is lazy by default and cannot be configured otherwise. This leads to excessive code just to get all the data needed for a use-case.

- PONO domain classes, must inherit from EntityObject base class. This also ties the domain classes strongly to EF, not very invasive.

- Rich domain model (MS encourages the Anemic Domain Model anti-pattern while rich models are supported through partial classes, which I also think is awkward.)

- Testability, to unit test reverse-engineered EF classes you must test against the DB at all time. This is simply to slow and inconvenient, and probably another reason for advocating the anemic domain model approach.

- Domain driven development, cannot start with my classes and their behaviour first. This also makes TTD less attractive.

- Seamless team collaboration. EF causes a lot of friction in team collaboration scenarios due to too many merge conflicts happen to frequent.

Resolving lazy-eager problems and enable to define a default fetch plan would be good starting points as well as getting rid of the EntityObject base class. Try to make EF as invasive as possible. Remember that the domain model is the heart of any application. Domain models should thus be pure and simple, if possibly PONO/POCO. Technologies around the domain model change frequently, and to be able to embrace change, these concerns should be separated away from the heart of your application.

Also, be careful with trying to solve all problems with Attributes. Now WCF also want you to endorse your entities with DataContract and DataMember attribtues. The code would read awfully if domain classes must worry about serialization and db-mapping issues. In stead of the well known XML-hell we might get a new Attribute-hell!

Cheers,

Steinar.

re: Vote of No Confidence

Andrey — Mon, 24 Nov 2008 12:55:06 GMT

I really appreciate the power of LINQ and EF as a former FoxPro programmer.

Great step, keep it up.

re: Vote of No Confidence

Mark Gordon — Wed, 20 May 2009 10:04:51 GMT

EDM/EF is a data technology abortion. Unforunately I stuck supporting an application written using this technology. I'm really not sure where your/MSFT head is at but obviously you are completely out of touch with the needs of your development community. The funny part is what you really need to incorporate into VS is a data centric language. The FoxPro core has been around 20+ years which MSFT owns the source. Microsoft's data technology you currently are developing last 1 or 2 years tops. Given you are completely clueless when it come to data why not steal ideas from something that works, the VFP model, instead of reinventing the wheel since you suck so bad at it. After all isn't stealing from apple how windows came to be.

re: Vote of No Confidence

George Bush — Sun, 09 Aug 2009 07:48:26 GMT

@ Mark Gordon: God, I thought you Fox Pro relics had all gone the way of the dinosaurs. Technology has moved on since the 1980s. You're ilk and also that WinForms lot have become outdated and no longer in demand thank god. Programming has become more of a profession that takes knowledge and the application of learned disciplines and skills, instead of the cowboy infested drag-n-drop from toolbox mishmash of imperative spaghetti code it used to be in the last decades.