A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
1–10 of 33 posts
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#2I appreciate the meme and get it as a response to everything being "serverless" recently. But I noticed some people started repeating it unironically, like cgi-bin was actually equivalent. I wish this project at least acknowledged it's a meme reaction. Maybe people these days don't remember anymore what cgi-bin really worked like? (or never worked with it?)
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#3Funny, reminds me of the whole cloud-to-butt thing.
On a different note, the whole serverless thing seemed weird to me until I looked into it an realized that it kind of is cgi-bin all over again.
I love it, someone else can manage and patch the server and worry about scaling. I just deploy my workload as a self-contained app or set of apps. Scale is only limited by the bank account.
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#4Funny, reminds me of the whole cloud-to-butt thing. On a different note, the whole serverless thing seemed weird to me until I looked into it an realized that it kind of is cgi-bin all over again. I love it, someone else can manage and patch the server and worry about scaling. I just deploy my workload as a self-contained app or set of apps. Scale is only limited by the bank account.
Also, the ever humorous millennials to snake people[1] and its ensuing news story[2]
[1] https://chrome.google.com/webstore/detail/millennials-to-sna... [2]https://gizmodo.com/new-york-times-issues-correction-after-e...
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#5I appreciate the meme and get it as a response to everything being "serverless" recently. But I noticed some people started repeating it unironically, like cgi-bin was actually equivalent. I wish this project at least acknowledged it's a meme reaction. Maybe people these days don't remember anymore what cgi-bin really worked like? (or never worked with it?)
I think you're the one mistaken, here. The interface for cgi-bin was that you created a process per web request. The interface for Serverless is that you create a process per web request.
The tooling is a lot better for Serverless, but I'd hesitate to call it 20 years better - I'd hesitate to call it 10 years better.
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#6I appreciate the meme and get it as a response to everything being "serverless" recently. But I noticed some people started repeating it unironically, like cgi-bin was actually equivalent. I wish this project at least acknowledged it's a meme reaction. Maybe people these days don't remember anymore what cgi-bin really worked like? (or never worked with it?)
I think you're the one mistaken, here. The interface for cgi-bin was that you created a process per web request. The interface for Serverless is that you create a process per web request. The tooling is a lot better for Serverless, but I'd hesitate to call it 20 years better - I'd hesitate to call it 10 years better.
> The interface for Serverless is that you create a process per web request.
That is not how it works in the most popular solutions. Both aws lambda and Google cloud functions will keep the server instance running after your first request and only dispatch function calls - you can cache data between requests in globals. That's also why the first cold request is so much slower (it actually loads the code for the first time).
Most serverless function environments are much closer to managed fastcgi, than old cgi-bin approach.
This is mentioned right in the service docs: https://cloud.google.com/functions/docs/concepts/exec#functi...
Also, thanks for proving my point - some people don't understand it's a meme, not real equivalence.
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#7Ok I’ll bite. The argument here is that serverless is similar to cgi (common gateway interface, the original way a web server handed requests to an executable) in that they both spawn a process for each ‘dynamic’ backend request (static files are served from a cdn in serverless model).
This is of course a false and misleading analogy - the reason spawning a new process for each request is bad is that it doesn’t scale on a single server, but in the serverless model the infrastructure serving this is scaled up exactly so you don’t have to worry about this aspect.
Response times are measurably very good for go python and js, in our experience, and aws serverless will try to reuse you running process for subsequent requests (warm start).
And most importantly — there are no servers to manage.
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#8Funny, reminds me of the whole cloud-to-butt thing. On a different note, the whole serverless thing seemed weird to me until I looked into it an realized that it kind of is cgi-bin all over again. I love it, someone else can manage and patch the server and worry about scaling. I just deploy my workload as a self-contained app or set of apps. Scale is only limited by the bank account.
It’s a really apt comparison. Before PHP and embedded scripting with HTML, and entire contained applications with MVC frameworks, we had our static HTML and our CGI.
Now we have serverless and HTML that talks to our cgi-bin using JavaScript. Life is full circle. Client/server back to the cloud, which reminds me of the dumb-terminal/mainframe days.
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#9Earlier quoted context omitted.
I think you're the one mistaken, here. The interface for cgi-bin was that you created a process per web request. The interface for Serverless is that you create a process per web request. The tooling is a lot better for Serverless, but I'd hesitate to call it 20 years better - I'd hesitate to call it 10 years better.
> The interface for Serverless is that you create a process per web request. That is not how it works in the most popular solutions. Both aws lambda and Google cloud functions will keep the server instance running after your first request and only dispatch function calls - you can cache data between requests in globals. That's also why the first cold request is so much slower (it actually loads the code for the first…
so it's not serverless then...
Re: A browser extension that replaces occurrences of 'serverless' with 'CGI-bin'
#10Ok I’ll bite. The argument here is that serverless is similar to cgi (common gateway interface, the original way a web server handed requests to an executable) in that they both spawn a process for each ‘dynamic’ backend request (static files are served from a cdn in serverless model). This is of course a false and misleading analogy - the reason spawning a new process for each request is bad is that it doesn’t scale…
> Response times are measurably very good for go python and js, in our experience, and aws serverless will try to reuse you running process for subsequent requests (warm start).
This is because most modern serverless implementations are closer to FastCGI (which had similar benefits over classic CGI). But the concepts are the same.