flyingpants562021-08-21 #
Is there a reason that I'm not aware of? Are only certain titles accepted?
Why only 13,000 titles for Windows, when Steam has over 50,000 titles?
flyingpants562021-08-21 #
Because editors follow their passions, and most of them prefer to update retro consoles/computers instead of modern ones, or Windows.
In my case, for example, I try to fully update the games I've played. But sometimes I get curious about some old console that I have never touched,
and I also work on its games.
In 2020 Steam released about 10'000 games, and for a small team like us, that never had more than ten editors at the same time, it is quite a job.
Bear in mind that there is no automatic bot that regularly scrapes other sites, most of the games are researched and entered manually.
There are some partially automatic task that clean-up, refresh the data, but they still require manual actions for every entry.
Anyone can join and help us. Even some occasional updates can really help.
In my case, for example, I try to fully update the games I've played. But sometimes I get curious about some old console that I have never touched,
and I also work on its games.
In 2020 Steam released about 10'000 games, and for a small team like us, that never had more than ten editors at the same time, it is quite a job.
Bear in mind that there is no automatic bot that regularly scrapes other sites, most of the games are researched and entered manually.
There are some partially automatic task that clean-up, refresh the data, but they still require manual actions for every entry.
Anyone can join and help us. Even some occasional updates can really help.
@Andrea
I'm also not a fan of automatic bot scrapings as this usually results in tons of data you don't want.
But what I'd really like to see in the future is a manual scrape bot to get information from main sites.
Like I create a Title and have the option to manually getting data from Steam/Nintendo/Microsoft/ etc.
It then asks me what Data i want to take over for example.
Because stuff like Title, Year, Publisher, Developer, Screenshots especially for new Stuff is just copy and Paste directly from the Big Stores.
I'm also not a fan of automatic bot scrapings as this usually results in tons of data you don't want.
But what I'd really like to see in the future is a manual scrape bot to get information from main sites.
Like I create a Title and have the option to manually getting data from Steam/Nintendo/Microsoft/ etc.
It then asks me what Data i want to take over for example.
Because stuff like Title, Year, Publisher, Developer, Screenshots especially for new Stuff is just copy and Paste directly from the Big Stores.
In theory, ChatGPT3 or the upcoming ChatGPT4 can be very good at processing scraped data and comparing it against the existing game entries at UVL, so that the majority of the extraneous data will be correctly excluded. And where it does make mistakes, it can be told exactly what it's mistake was and it is unlikely to repeat that same mistake ( if the correction is phrased correctly ). Of course for that level of training you would need two things. Premium (paid for) use of ChatGPT3 and a team of trainers that understands how ChatGPT3 thinks. Note, while ChatGPT3 can provide suggestions on the best way that a trainer should correct it, it does not fully understand itself well enough that this tactic is a reliable way to train trainers. (Or so chat GPT tells me.) When it comes to suggesting how trainers can best correct it, ChatGPT3 can only offer an alternate perspective on the best way, not necessarily a correct one.
Also note, when I talk about chatGPT as if it's self-aware, I merely imply to mean that it simulates certain aspects of self-awareness in many situations. On a technical level it doesn't "understand" anything, but you could ask it about fictional subjects and it will give you information about that subject in the context that is fiction (as if it understands the difference between fiction and reality).
I asked ChatGPT3 specifically about this several different ways and dozens of times. I gave it some sample web pages to look at from UVL and seven other game database sites. And ask for an estimate to compare the seven sites and enter or correct 75,000 games using the UVL API. Estimates range from 268 days exactly to 281 days and 4 hours and 8 minutes and 31 seconds for a specifically trained version of ChatGPT3. ChatGPT4 would take longer but be more accurate.
Also note, when I talk about chatGPT as if it's self-aware, I merely imply to mean that it simulates certain aspects of self-awareness in many situations. On a technical level it doesn't "understand" anything, but you could ask it about fictional subjects and it will give you information about that subject in the context that is fiction (as if it understands the difference between fiction and reality).
I asked ChatGPT3 specifically about this several different ways and dozens of times. I gave it some sample web pages to look at from UVL and seven other game database sites. And ask for an estimate to compare the seven sites and enter or correct 75,000 games using the UVL API. Estimates range from 268 days exactly to 281 days and 4 hours and 8 minutes and 31 seconds for a specifically trained version of ChatGPT3. ChatGPT4 would take longer but be more accurate.