The reason I bring this up is because this page uses code directly from my blog post (the function print_graph_nodes). I was surprised to see that on this page, although there is no reference to my post here.
I remember wasting time on writing this function to ensure that the formatting (ie the '=======================INPUT===================') looked good :P This had actually helped me find a bug in my own code.
Happy to know I helped in some way :)
Anyone who's seen the great Christopher Guest documentary on the subject would agree. ;)
catch (Exception e)
If Java requires you to catch the exception or declare a throw, then it seems like it should throw until you get to an outermost stack frame that's sane enough to handle it (open an error dialog box, retry, log the error, etc).
Sorry, the app and the idea are super cool, and keep up the good work. This is intended as constructive criticism.
with something like
Log.e(TAG, "Caught and ignored an exception", e)
I've used it to identify plants and insects when on walks. Really fun to use.
Feels very Star Trek.
I'm more interested in the wine pairings.
The binary outcome consequences of misidentifying a dog breed in an app are not really important, but the ability for the breed to be a factor in informing some other aspect of a model is useful to businesses that optimize cash flows.
For example, I love stafforshires, and I also recognize that a machine learning scheme that added them as a factor in a credit risk evaluation might disadvantage a number of people, but be more profitable than a model that did not include it. If this is offensive, consider that breed specific legislation already allows you to be denied home insurance (and consequently, mortgages and equity loans) based on having one of these dogs today.
The code in this project could be used to mine social media images for breeds correlated with credit risk factors, and offer interest rates for the owners accordingly, and a credit instrument portfolio that incorporated this factor could outperform one that did not. In regions with breed specific legislation, someone with an banned dog has already demonstrated a deviation in their risk behaviour.
Shelters are not full of failed show dogs. In fact the majority of them are mixes, particularly mixes preferred by people who were guided by sentiment over more sound reasoning factors when they got the dog in the first place - hence why the dog is in the shelter, and why they probably can't get credit on favourable terms.
Is this just or ethical? No, it's predatory, but that's the social media bargain, and projects like this are worth keeping track of if you like dogs, and access to credit.
Many also pick up pets because of child pressure or simply fads. 101 Dalmatians was a shit show because of the number which ended up abandoned as people didn't realize that a large dog was an impact.
Rescue groups are full of dogs that are abandoned because they are too old, required expensive medication, developed eye issues, or worse. As in, far too many people are not mature enough to own pets. Sadly the same could be said for many having children as well.
If I were king of Disney or Youtube I would ban all cute puppy footage.
I'd also discourage anyone who has never owned a dog from purchasing a particular breed. If you have no experience with dogs, get a mutt. If you want a larger dog, get a lab-Shepard cross. If smaller, any of the bichon-havanese crosses.
(spoilers, of course)
We see this flipped in pg-13 movies these days where writers, who grew up watching dogs suffer under the old rule, wouldn't dare kill off a dog (independence day, 1 and 2). So the presence of a dog in a scene now reassures the audience that all will be well. But in an R-rated movie... if I see a dog on screen I'm tempted to walk out of the theater because I know it won't end well.
People do want specific traits in dogs though, whether for hunting, for companionship, or agility etc. I just wish breeding dogs had better oversight and certification along with better custodianship of the breed's viability. (i.e. none of this more and more extreme visual traits that many "breed experts" judge against)
Especially when you understand just how sentient these beings are. Euthanizing them shouldn't be so widespread.
We have serious responsibilities to another species that we are not maintaining, we should be doing better for our dogs.
In the UK we have so many staffordshire bull terriers in kennels, or abandoned, they get 24 hours when caught, and then euthanised if not claimed by owners or a rescue, in my opinion it is a product of both irresponsible breeders and irresponsible owners.
We are currently fostering a staffie, he is the sweetest thing, a little snugglemonster, but he is one of the lucky ones, most don't survive the 24 hours.
Demand definitely has something to do with it.
I would also argue that quality of supply has something to do with it too. There are good breeders and bad breeders.
I tend to think that the good breeders make every effort to avoid allowing health issues to get bred to the next generation. My breeders would take a dog with health issues out of their breeding program the moment they found out. It also helps that a large number of the people who got their dogs from my breeder have gotten to know each other on social media, so if there's an issue, we're much more likely to hear about it.
I have a suspicion that bad breeders likely outnumber good breeders, and produce much more of the supply of purebreds, which increases the incidence of purebreds with health issues.
My breeder has a waiting list, and you are not guaranteed to be accepted as a potential adopter (i.e., there's a detailed questionnaire to fill, and you have to pass their "smell test" when they meet you). They have a fixed number of litters per year (litter size can vary though, so you can wait a long time to get a dog), and they never increase that number no matter how much demand there is.
They never have a problem finding a home for their pups, because they have a reputation for being good breeders and for their dogs not having health issues.