Disclaimer:
I'm unsympathetic. I guarantee you all I am far more sick of reading this post than you could ever be. I've read it and reread many many many times. 
But this needed to be addressed with details.@Neko Rirenil: It will keep growing until you can never finish it.
@Volki: I had some more to say about the rest of your post Volki, but I'll leave it at this: at least have the decency to not strawman what I said if you really insist on debating. I have no reason to debate your inventions on what I am saying and so far you are largely too focused on making up weird extrapolations to pay attention. Now, I'm going to focus on this dog thing for rest of the post and keep in mind that
the length of this post is your fault. You started this.
Here's the rub with the dog breeds. Lets start with the basics.
Introduction:People who say it takes a long time never disputed the rise of diversity. They are pointing out that diversity is too small a step to quantify it as evolution. Yeah, they say it plays a role, but don't stretch the concept too far. No one is debating whether dog breeds show different traits, but they will never truly be anything but the same species for ever more unless there is enough change to the genes. That takes time without directly tinkering in the genetic code. If the set of genes contain alleles for light brown fur and black fur, you have done nothing special by selectively breeding for the black trait as you never eliminated the brown trait from the genes. The brown trait may not be often expressed, but it is still a part of the series meaning no net change. You haven't done terribly much more evolutionarily than someone that just dyes their dog pink with food dye. This is why breeds can still have members that have traits other than the desired ones you didn't eliminate the other possibilities. Now here is one thing to consider on this topic, apparently some dog breeds have lost genes even though they are all still able to interbreed. Interbreeding is why they are considered one species, but there is an issue with how the dogs come to loose genes. I'll get to what's wrong with that loss and why it's not evolving dogs a little further down.
Let's work this through logically. Evolution is supposedly the process that lead to the rise of new species, and it is supposed to have done this by the passing on of useful genetic traits to the next generation which let them survive better than others of the species. Superficially selecting sets of preexisting traits without a
sustainable genetic component goes against this model. Those dog breeds may not revert to wolves, but they can't pass on a self sustaining uniqueness; at best they still require deliberate human intervention. They and their descendants will always be the same species as wolves until you produce a large enough, again sustainable, change in genes which does not happen quickly without specifically tinkering in the genetic code. That is why scientists say evolution takes time. It's a heck of a lot more complex than selectively breeding in snoopy spots.
Now, that being said, this post took so long is because I decided to dig into dog genetics research--the things I do to make a point -_- --and what that means for breeding as evolution and here's what I got. First, I found that people like to dance around and state things but do not like to quickly get into the important info at the heart of the issue unless you look at scientific papers. This is frustrating since that meant several searches were less than informative. But I got somewhere eventually.
The Research:Let's start with some defintions for reference. You can double check your own sources if you like or skip if you already know this stuff:
Polymorph: when two or more clearly different phenotypes exist in the same population of a species — in other words, the occurrence of more than one form or morph.
Phenotype: the set of observable characteristics of an individual resulting from the interaction of its genotype with the environment.
Genotype: The entire set of genes in an organism.
Allele: An allele is an alternative form of a gene (one member of a pair) that is located at a specific position on a specific chromosome.
Genetic Bottleneck: In which the genetic variation within a population and the potential to adapt to a changing environment decrease.
Relaxation of selective constraint on dog mitochondrial DNA following domesticationNow you can find bits of specifics here and there, but here is a paper that actually gets some of the more useful information in one spot. There is a lot of stuff related to methods that someone in that field would better understand, but here's what I got out of it:
Dogs are in fact polymorphs of each other and the breeding that went into creating specific breeds produced a lot of differences in genes in the breeds; however, it also caused lost genes producing genetic bottlenecks by things such as inbreding. Further, the majority of the current dog breeds were not the products of the long history of humans breeding dogs but rather they were developed in the 19th century. This means that most dog breeds are a far more recent occurance than I was previously thinking. now one thing that I wish I understood better about this paper is what they mean by wide difference in genes. Do they mean the individual genes are different from each other or are the genes they found just in occuring in a variety of different partterns. I suspect it is the later and I suspect that if you understand the methods part better that you'll know what they mean clearly. Oh well.
These findings may indicate that most breeds were somewhat homogeneous prior to being officially recognized by a breed registry. Indeed, breed clubs have to demonstrate a well-documented history and a well-described conformation standard prior to recognition of their breed by a registry. However, forces such as founder effects and bottlenecks (as a result of popular sires, severe changes in population sizes, and intense phenotypic selection) will continue to contribute to a decrease in genetic diversity after registry recognition.
Now I believe what they mean by homogeneous here is with reference to allele occurance. But here we get into why I think dogs were said to not be evidence of evolution. Yes humans have forced selection here, but it's not producing viable genetic pathways, but rather genetic bottlenecks. You are reducing genetic diversity. As I said earlier, the whole evolution bit relies on passing on useful adaptive traits, but the introduction of these bottlenecks work directly against that. It drops genetic diversity. These breeds may allow for artifical preservation of weak genes, and are still dependent on Humans to maintain breed purity. This makes a bad case for dog breeds as an example of evolution, not even artifically aided evolution.
Now this of course begs the question, "is this really an issue?" and like it, "How common are genetic disorders in dogs?" So let's explore that. Here's a paper:
Prevalence of inherited disorders among mixed-breed and purebred dogs: 27,254 cases 1995–2010)Wow... now this was an interesting paper and not at all what I was expecting. There is a really nice, neat summary of relevant info at the beginning but I'm going to pull some more than that too. But here is the nice and pretty of it:
OBJECTIVE: To determine the proportion of mixed-breed and purebred dogs with common genetic disorders.
DESIGN: Case-control study.
ANIMALS: 27,254 dogs with an inherited disorder.
PROCEDURES:
Electronic medical records were reviewed for 24 genetic disorders: hemangiosarcoma, lymphoma, mast cell tumor, osteosarcoma, aortic stenosis, dilated cardiomyopathy, hypertrophic cardiomyopathy, mitral valve dysplasia, patent ductus arteriosus, ventricular septal defect, hyperadrenocorticism, hypoadrenocorticism, hypothyroidism, elbow dysplasia, hip dysplasia, intervertebral disk disease, patellar luxation, ruptured cranial cruciate ligament, atopy or allergic dermatitis, bloat, cataracts, epilepsy, lens luxation, and portosystemic shunt. For each disorder, healthy controls matched for age, body weight, and sex to each affected dog were identified.
RESULTS:
Genetic disorders differed in expression. No differences in expression of 13 genetic disorders were detected between purebred dogs and mixed-breed dogs (ie, hip dysplasia, hypo- and hyperadrenocorticism, cancers, lens luxation, and patellar luxation). Purebred dogs were more likely to have 10 genetic disorders, including dilated cardiomyopathy, elbow dysplasia, cataracts, and hypothyroidism. Mixed-breed dogs had a greater probability of ruptured cranial cruciate ligament.
CONCLUSIONS AND CLINICAL RELEVANCE:
Prevalence of genetic disorders in both populations was related to the specific disorder. Recently derived breeds or those from similar lineages appeared to be more susceptible to certain disorders that affect all closely related purebred dogs, whereas disorders with equal prevalence in the 2 populations suggested that those disorders represented more ancient mutations that are widely spread through the dog population. Results provided insight on how breeding practices may reduce prevalence of a disorder.
OK, so the conclusion tells the some of the story. Some of the disorders are more prevalent in certain purebreeds and related breeds, and some were found equally between purebreeds and mutts. Apparently the occurance of the issues in either mixed or purebreds depended on the condition. But I do find the combined amount of genetic disorder between the two unfortunate. It was roughly 30% of the dogs in the study; there were a little over 90,000 dogs' records checked. Quick note, if you want to understand the limitations in deciphering the data you can go read the discussion section of the paper.
Dogs are second only to humans in the number of hereditary diseases identified in the population.
A quick googling says roughly 350 genetic disorder for dogs. :/ Apparently with great diversity comes great diversity of problems. I'd say that this starts to answer the question of is it really an issue, but I will concede here that I do not know the relative prevalence of genetic issues in the animal kingdom.
It is likely that with breed refinement for specific tasks and morphology, some mutations accompanied selection for those traits. Rigorous selection pressures to refine the breeds by inbreeding and bottlenecks would contribute to a loss of genetic diversity, thereby increasing the likelihood of recessive disorders within a breed population.
This section was talking about proposed reasons why purebreds might be seen as more prone to disease, but what I wanted to point out here is that it specifically references inbreeding and genetic bottlenecks. Now we all know that dog breeds can be inbred but this is a clearer statement that breeding practices also produce genetic bottlenecks which could lead to increased occurrence of recessive genetic issues. That means that it is more likely that mom and dad dog are likely to have the genes for a disorder that could be prevented in the pup if only one parent had the gene for it. There were two citations in this section as well, didn't follow those either for the record. But it's those genetic bottlenecks that are noteworthy in this issue of dogs as evidence of artificial selection.
Of the 24 disorders assessed, 13 had no significant difference in the mean proportion of purebred and mixed-breed dogs with the disorder when matched for age, sex, and body weight (Table 2). ...
In contrast, 10 disorders were more prevalent in purebred dogs, compared with those found in mixed-breed dogs. ...
This is just a more specific breakdown of the findings, but I do find it sad that in 13 disorders, essentially both kinds of dogs were susceptible. Now there are a few reasons why this could be, but the paper kept coming back to common ancestors carrying the trait and small breed gene pools due to low amount of individuals to breed. Now bear with me here, if you have few individuals that fit an ideal you want for a breed, and you try to create a whole breed from them, that presents a potential problem with a low gene pool. With that scenario, you are setting up for a problem right from the beginning.
Disorders may be associated with breed derivation or with breed bottlenecks. Such an example is the Irish Wolfhound, a breed with relatively few dogs registered annually. In the mid-1800s, the Irish Wolfhound underwent a population bottleneck so severe that the breed was thought to be extinct. The reduced effective population size suggests a relationship with the concomitant increased risk of dilated cardiomyopathy in Irish Wolfhounds. Indeed, as many as 1 in 3 Irish Wolfhounds may be affected with this disorder.
The disorders they are talking about in the first sentence are the ones associated with a purebreed. What this is saying is genetic bottleneck or a breed being related to a breed with this predisposition is likely the reason you can more or less expect this issue in that breed. Now if creating dog breeds is contributing decently to producing genetic bottlenecks, it's also increasing the risk of non useful or even outright harmful genes.
Conclusion:Now let's just skip along to the point I want to make about dog breeding and evolution. Now if we take the definition of evolution I stated earlier: "passing on of useful genetic traits to the next generation which let them survive better than others of the species " not the colloquial definition of "its changed", and we compare this to the information in the two papers, and please feel free to go through the papers and make a counterpoint here, then the rise of dog breeds doesn't count.
One, the process by which we selectively breed for traits to make current dog breeds is actually contributing to genetic bottlenecks which run counter to the supposed process of evolution. No not every dog is broken hipped and having heart attacks, but these genetic disorders often seem to be connected to traits that we have been selectively breeding for in various dogs. Even in the last paper they speculated that some of the disorders found in both mutts and purebreds could come from a common ancestor and,
Perhaps the same desired traits that made dogs a favorable species for domestication were linked to alleles for hyperadrenocorticism, hypoadrenocorticism, cancers, hip dysplasia, lens luxation, and some cardiac disorders that were not found to be different between purebred and mixed-breed dogs.
Pointing to selective breeding possibly bringing along the genes for those disorders
Two, the genetics for the dog breeds we have seems to be questionably sustainable. Even if a dog breed is somewhat unique, it looks like the more you selectively breed for a trait the more risk you run of also trapping harmful genes in with the desirable trait due to things such as inbreeding, low population to begin with, and not checking the genetics of the starter dogs carefully enough. You accumulate enough harmful genes and you've shot that endeavor in the foot entirely.
Three, this process started with human intervention and it seems will require humans to stay on it to maintain it. Without humans to maintain pure breeds, dogs should just start to blend again. After all, for a dog, one mate is as good as another, so gains in isolating breed traits should start to fade without intervention. Breeding is also typically not done with trying to create an evolutionary pathway in mind, so more than likely breeding as we do it now shouldn't produce anything but more of the same species. Supposedly it's the environmental pressures that will shape that process of evolution, but we know that humans are artificially tampering with that environment. Humans skew the whole picking off less viable members by keeping dogs as pets, even some with severe genetic issues, and in some cases they breed dogs past a safe point in terms of producing genetically viable offspring.
All of this seems to work against standing definitions of evolution. Dog don't really count. They can't just due to how it's defined.
