Language:
switch to room list switch to menu My folders
Go to page: First ... 68 69 70 71 [72] 73 74 75 76 ... Last
[#] Mon Jan 17 2011 12:24:06 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

There is a reason why protocols should be mostly human-readable:
people have to develop in them.

People should use debuggers.

It took me a while but I've finally caved on the everything should be binary.
Now I'm in the camp of: client commands to the server should be ascii but data should be binary esp if there's a lot of it. And http fits that pretty well.

[#] Mon Jan 17 2011 14:34:38 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

It's much easier to troubleshoot a protocol that allows you to telnet to the port and type commands to the server. HTTP is mostly a win in that regard.

And yes, IMAP succeeded in terms of being the standard for client/server access to mailboxes. Pretty much every fat-client other than Outlook uses it. Believe it or not, the whole world has *not* migrated to webmail.

[#] Tue Jan 18 2011 14:13:54 EST from LoanShark @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

there are wireshark pulgins to make them human readable.

Clearly this discussion is about the implementation of new protocols as well. Do you really want to have to write your own wireshark plugin just to have a chance of understanding something you designed yourself?

[#] Tue Jan 18 2011 14:14:31 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Now I'm in the camp of: client commands to the server should be ascii

but data should be binary esp if there's a lot of it. And http fits
that pretty well.

+1

[#] Tue Jan 18 2011 16:03:18 EST from dothebart @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

 

Di Jan 18 2011 14:13:54 EST von LoanShark @ Uncensored Betreff: Re:
there are wireshark pulgins to make them human readable.

Clearly this discussion is about the implementation of new protocols as well. Do you really want to have to write your own wireshark plugin just to have a chance of understanding something you designed yourself?

give wireshark the asn1 description, and you're done. Don't know whether you need to recompile or not, but most of wiresharks protocol analyzers are done with asn.1



[#] Tue Jan 18 2011 20:55:21 EST from LoanShark @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

give wireshark the asn1 description, and you're done. Don't know
whether you need to recompile or not, but most of wiresharks protocol
analyzers are done with asn.1

ASN.1 is a parse tree, not a sentence. Even with that, it's not going to be as human-readable as a DSL.

[#] Tue Jan 18 2011 22:17:08 EST from IGnatius T Foobar @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

What's with the sudden flurry of ASN.1 fandom? ASN.1 is the primary reason why the "L" in LDAP is a pathetic joke.

[#] Wed Jan 19 2011 15:42:23 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

sad actually.

[#] Wed Jan 19 2011 15:56:42 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

sad actually.

?

[#] Wed Jan 19 2011 18:14:45 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

No test or stage environment, just throw new code at some of the users and see what happens.

Not surprising.

Of course, the article mistakenly refers to the site itself as the company's product. The site is not the product, and its users are not the customers.
The users are the product, and its advertisers and other revenue sources are the customers.

[#] Thu Jan 20 2011 16:35:55 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]


... and the MS investment valued those users at $300 a pop.

[#] Thu Jan 20 2011 16:37:16 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

No test or stage environment, just throw new code at some of the users

and see what happens.

A/B testing is -very- important for a site that has to generate revenue, but that's not the way to do it.

If you read further down the article, there are some contradictory claims: Facebook doesn't QA! No wait, yeah they do! *shrug*

[#] Fri Jan 21 2011 15:17:51 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

Jan 19 2011 3:56pm from LoanShark @uncnsrd
sad actually.

?



Well as you point out, that guy was guessing a lot and a lot of what he guessed at was wrong, I wouldn't take too much value in the 'facts' in that post.
I added a comment, but basically the sad is that for all of their zillions of developers whether they throw things over the wall or go through some kind of development process, they're all php programmers, and I can't have a hell of a lot of respect for that.
There's a few guys working on scalability and the javascript munger but I read somewhere that the vast majority of fb programmers are php guys. And that's why they made the php compiler rather than stop and reimplement everything in a real language.
I give them credit for their scalability, which now that I think about it is probably done by mass duplication of data rather than any smart indexing scheme or something like that and lots of hardware thrown all over the place.
I have a lot of not nice things to say about google, but I have never questioned their technical ability.

For all of these 500 programmers throwing shit over the wall every few days... where is it? At this rate there should be a few hundred thousand features or tools or functions or something. I dont' see it.

[#] Fri Jan 21 2011 15:18:17 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

... and the MS investment valued those users at $300 a pop.

I'd guess that MS office users are valued at more than that. They should invest their money in microsoft. :-)

[#] Mon Jan 24 2011 08:26:25 EST from IGnatius T Foobar @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

For all of these 500 programmers throwing shit over the wall every
few days... where is it? At this rate there should be a few hundred
thousand features or tools or functions or something. I dont' see it.


And that's why I am not buying into the whole "Google is not the cool place to work anymore, all the hot talent is going to Facebook" hype. Facebook is huge but it isn't really cutting-edge. If you're a genius with a PhD in complex stuff engineering, why would you go looking for work at what basically amounts to a giant BBS written in PHP?

Netscape at one time made a transition from a place where people went to build something great, to a place where people went to work because it was already great. If Facebook has already made that transition then it does not bode well for the company. At least Google has a business model.

[#] Mon Jan 24 2011 22:21:35 EST from Ford II @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

yeah I heard that too and it sounded a little fishy.
Probably google doing that cut-the-bottom-5% thing.

[#] Wed Jan 26 2011 09:56:26 EST from skpacman @ Uncensored

Subject: Re:

[Reply] [ReplyQuoted] [Headers] [Print]

Recently, I was forced to realize the importance of a collation call-out (or lack thereof) in a MSSQL query in PHP.......



[#] Thu Feb 10 2011 00:19:58 EST from LoanShark @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]


Ok, my current project requires...

[drum roll]

EBCDIC.

Yea, definitely should have stayed put ;)

[#] Thu Feb 10 2011 10:29:30 EST from skpacman @ Uncensored

[Reply] [ReplyQuoted] [Headers] [Print]

For those that dont speak in acronyms:

EBCDIC:  http://bit.ly/CCC34



Go to page: First ... 68 69 70 71 [72] 73 74 75 76 ... Last