[gopher] Re: Mozilla bugs about Gopher, and a dangerous one

[Chris <chris@xxxxxxxxxx>]

Ok I'll try and clarify a few thing,
by the way go get a cup of coffee first :)


On Mon, 16 Apr 2007 10:13:24 -0700 (PDT)
JumpJet Mailbox <jumpjetinfo@xxxxxxxxx> wrote:

> Sorry Chris, I wasn't alluding to "Local Veronica" (or V-2, or even Jughead), 
> all of which are excellent tools.  I should have been more clear in that I 
> was refering to the no longer available non-local Gopherspace-wide Veronica, 
> such as what was once run on the now defunct UMN Gopher.  When I said "dead", 
> I meant that the software is not running on Gopherspace, and therefore there 
> was no need to make accomodations for it (and therefore as well, should it be 
> brought back again, it also doesn't have to meet any pre-existing parameters).

I am running the original "non-local Gopherspace-wide Veronica"
(I added deleted parts from local-veronica software to make it back into the 
original non-local-veronica)
It is running and well. Depending on spidering between 1.2 million selectors at 
a high and 678,000 selectors at a low.
Also I run a "very large section of gopherspace jughead" as well as a local 
jughead and a local veronica as well as a few other network tools based on 
these. And some other less obvious to the world search engines housed on the 
net.

>    
>   ==========
>    
>   I am however wondering why the Port issue is such a hot spot?  Yes, I do 
> understand that in regards to Ports, NONE of the Ports should be "off 
> limits".  If I want to run my HTTP server on Port 21 (the common FTP Port) 
> instead of Port 80 (the common HTTP Port), my software should not prevent me. 
>  
> 

Right I agree none of the ports are "off limits" rules that would do this are 
imo bad.
further explanations below
   
>   Still, there are indeed several advantages to having software stick to 
> specific Ports.  The biggest advantage is the ease in configuring FireWalls.  
> By restricting the number of Ports open to the world (and it always mystified 
> me why they made so many Ports anyways, 512, or even as little as 256, whould 
> have probably been more than sufficient based on real-life Port useage 
> statistics) Network Administrators can more easily detect and prevent abuse 
> of their Network.  
>
I disagree, 256? never work, thats not enough, not even close.
    
>   Another advantage of having certain Protocals stick to a single (or at 
> least a very limited range) of Ports is the ease of which a Client / Browser 
> can detect a Server.  This is why there are millions of HTTP web pages 
> operating on Port 80 (or Port 8080) rather than on Port "X".  Non-technical 
> persons who surf the Web, typically don't want to be constantly appending 
> Port numbers to the end of an IP address.
>

Ease to detect? Use a searchtool that sees them and a client that does as well 
and you dont have to append anything if it is a link or is in a search engine.
    
>   The standardization of Port numbers has been felt so important in fact, 
> that when you speak of a Protocol, you typically assume a specific Port (HTTP 
> = 80, FTP = 21, SMTP = 25, DNS = 53, etceteras).  IANA even maintains a list 
> of what they call "Well Known Port Numbers" : 
> http://www.iana.org/assignments/port-numbers 

Please read rfc1436 on your own server and in paticular pay attention to 3.2 
Server portability and naming
3.5  Building clients
3.1  Locating services
and:
Note: Port corresponds the the TCP Port Number, its value should
      be in the range [0..65535]; port 70 is officially assigned
      to gopher.
And please read carefully what Jon Gorzen and others wrote at mozzilla.
Bug 158888
and follow back to:
Bug 71916

I'll add more later in this email.
But "There is nothing to prevent secondary servers or services from running on 
otherwise named servers or ports other than 70"

>    
>   With only about 200 or so Internet Gopher Servers worldwide (and only about 
> a dozen still actively updated), why is using Port 70 such a hardship???  
>   Is it because of a imperative need to run Dual Protocol servers?  (GN will 
> let you serve the same content on two different Ports simultainiously, Port 
> 70 and 80).

Dual? If only.....
  
>   Is it because a Legacy Server is already operating on a non-standard Port?  
> (Contact the Servers Administrator and ask them to switch it onto Port 70).  
>    
I am the Servers Administrator, and I'll try and explain more clearly below 
why, and it isnt "non-standard".

>   It is indeed unfortunate that many Browsers that understand Gopherspace can 
> not recognize Gopher on a Port other than 70.  This is due primarily to 
> lazyness on the part of the software programmer (i.e., many Browsers such as 
> older versions of Internet Explorer).  However, even many of the older 
> dedicated Gopher Clients could only view Gopher on Port 70 (test this for 
> yourself by downloading and trying some of these older clients: 
> gopher://home.jumpjet.info/11\Begin_Here\Clients ).  
It is not only a unfortunate it's wrong, particularly to redirect to another 
site.
Looking through your list of Unix clients and having run most of them I am not 
seeing these "port 70 only" clients your refering to, they wouldnt have worked 
if they didnt do the search engine ports and would not have been used.


>    
>   With such a limited selection of Clients / Browsers able to choose a Port 
> other than 70, why not just stick with Port 70??  Should a unique Server have 
> to operate on a different Port, the Server Administrator should just assume 
> that his Server will be of limited accessability as his patrons must 
> INTENTIONALLY use one of the very few Clients / Browsers that can operatate 
> on a Port other than 70 (not that undue a burden, as persons operating, for 
> example, a FTP server on a Port other than 21 also face the same issue in 
> regards to limited Client choices for their patrons). 

You have an ftp client that cant do other than 21?!? 
>    
>   So, in conclusion; Why is operating a Gopher Server on JUST Port 70 such a 
> burden?

Gopher "search engines" are servers, they run on their own ports, ports such as 
the 3000, 2347 etc etc. If they ran on 70 they could not (easily) be run on the 
same machine right? So what you are suggesting is that I run all my gopherd and 
pygopherd servers on 70 as well as all the search engines (between 6 and 11 
depending on what I am doing) as well on 70 and then let the html and wap go to 
80 along with the various web servers and domain names.
Ok fine, lets look at it.
How? Well lets see, I could buy a block more IP addresses and give or take add 
another 6-12 boxen to my hardware, then set up my named to name serve these 
boxen with names like Veronica_Local.hal3000.cx, Veronica1.hal3000.cx, 
Veronica2.hal3000.cx, Jughead_Main.hal3000.cx,
etc etc etc...but wait, who's paying for this?
Me , so not such a good solution....
How about something like pygfarm or a dictonary server, but then you have for 
one, alot of scripts to write and for two added overhead and pipes and security 
issues,(assuming I would be using the original veronica or jughead to create 
and serve and then put into a database to farm out on a port 70 gopher page) so 
gosh seems like a lot of work to me to serve the gopher community a million 
selectors over several very wonderful original old time softwares that already 
work as they were intended to.
My gopher is on port 70, I serve other gophers not on port 70 and the owners of 
those don't care if they are seen publically (but I Know it irks them that in 
firefox, they see mine instead of theirs).
Veronica is "a secondary servers" per RFC1436.
this is clipped from a part of Jon G's moz bug post
**********************************************************************
RFC1436 specifies the Gopher protocol, and in section 3.2 asserts that "There 
is nothing to prevent secondary servers or services from running on otherwise 
named servers or ports other than 70".Further, it indicates that "The client 
software can locate and retrieve any item by the trio of selector,hostname, and 
port."
***********************************************************************
By the way telnet to a Jughead or Veronica with a query, it gives you a result, 
it is an independent secondary server serving a request independent of the main 
gopher.
So... with all that stated, yes its allowed but as to why I am not currently?
It's not so easy, and with people feeling that I could just switch them to 70 
and then saying heck firefox is fine as is, your effectively going to cut out 
anyone with a similar situation as mine from being able to offer up this free 
service. "Not so easy" may be an understatement.
I'm just a guy with 16 computers at my house, I only have so much time and 1 IP 
and a dsl line thats static which I currently pay triple the standard home user 
rate to have static.
So my "burden"? I think you can see the complexity.
It certainly isn't cause I am some old ludite saying no no no or anything so 
silly as that as a matter of fact I did ask Cameron long ago for the source for 
V2 but because of its complexity and the system he runs he decided he cannot 
offer the source up, I would be delighted as hell to offer my search engines on 
70 and would for legacies sake run them where they are at the same time, if it 
was something I could manage with a few bucks and a hundred hours, but the 
scope of it is larger than that and finally, all this cause moz is broken?
Nah I can't justify the time and or money to my wife and kids , they would 
certainly cut my (BIG) bundle of power cords :)
This isn't meant as a flame or a rant, I am just trying to explain why, V2 is a 
great great tool, and it's goal is to achieve what veronica did long ago, and 
my goal is to keep pace with V2 running what V2 tries to emulate and surpass, 
kinda funny really and it's been a grin truely, but as for me somehow just 
plunking my old Veronicas and Jughead all at the same time on 1 IP to 1 Port 
along with the gopher main...it just isn't going to happen easily , and just so 
you understand I have been trying to see how I might manage it, I never say 
anything "can't be done".
It just isn't going to be easy and I will _Never_ stop serving on the old ports 
either, I would only do both, with the original ports being primary.
At the same time, I have been working with another gopher fan on a small gopher 
proxy as well as an html rendering engine that would use perl GET as the 
fetcher and variations on this concept, this is of course for FreeBSD and 
GNU/Linux and similar O.S.'s I dont have any M$ here. But we are a long way off 
at the moment.
Hope this clears up the "Burden" and "Hot Spots" and gives you some
insight as to what some of the issues are. Currently as for a search engine 
searching all of gopherspace on port _70_ there is _Only_ V2,
mine are the _only_ others to even attempt it at the moment on _any port_.
(this is to say there are only 2, V2 and mine that have recently been updated)
If you find more send me the link as I have just re spidered and saw 
many sites leave for the second spidering in a row (results at 
/Links/Gopher/All_Active_Servers).
 
Regards,
Chris
hal3000.cx et all

>    
>   Yes, continue to hound Client / Browser software programmers to allow their 
> software to operate on more than one Port (if only because the software 
> SHOULD allow this flexability); but at the same time operate our handful of 
> Internet Gopher Servers on just Port 70.  
>    
>
   
> Chris <chris@xxxxxxxxxx> wrote:
>   
> Actually Veronica is not dead, it's alive and well, right on the port it was 
> intended to be on.
> It's port is quite appropriate for any millenium, any web browser which can't 
> handle ports is frankly past broken. How Moz acts, with it redirecting the 
> client to a different site while keeping the port looking as if it's where it 
> should be is quite unacceptable.
> Historically speaking it is un-questionably proper as it is.
> For the new millenium, why would you not allow the ports be used as they 
> always had?
> Veronica-2 is very nice, and is a valid tool, just as is Veronica.
> Chris
> On Sat, 14 Apr 2007 16:15:49 -0700 (PDT)
> JumpJet Mailbox wrote:
> 
> > I disagree with your reasoning about port 2347. Veronica is DEAD. I feel 
> > that doesn't matter what Veronica once did or did not do. A revived 
> > Veronica, Veronica-2, or any other future Veronica version should now run 
> > on whatever port is the MOST APPROPRIATE FOR THE NEW MILLENIUM. 
> > 
> > In the 10 or so years that there has not been a Veronica, the world has 
> > moved on. We now know what works best, and what should have been done 
> > differently. Here is a unique opportunity to do it correctly, and fix our 
> > mistakes (if mistakes they indeed were). If it happens to be port 2347, 
> > great. If its port 70, thats great too. If its some other port entirely, 
> > thats Ok as well. What is most important is that the port chosen is THE 
> > BEST PORT FOR THE JOB. 
> > 
> > V-2 works great for me on its current port 70. I know that much of the 
> > software currently available (be it a modern Web Browser, or ancient legacy 
> > Gopher Client) has trouble if Gopher is not on Port 70. Perhaps port 70 IS 
> > the best port for a Veronica??? 
> > 
> > Cameron Kaiser wrote:
> > > As for having the search engines stick to port 70.
> > > First I disagree, thats not a gopher problem its a moz problem .
> > 
> > I agree with this also. :2347 was always the standard Veronica port. In 
> > fact,
> > it's non-standard of *me* to run V-2 on the regular port and I've even been
> > mulling over putting a mirror onto port 2347.
> > 
> > 
> > 
> > ---------------------------------
> > Ahhh...imagining that irresistible "new car" smell?
> > Check outnew cars at Yahoo! Autos.
> > 
> > 
> > 
> 
> 
> -- 
> Join FSF as an Associate Member at:
> 
> 
> 
> 
> 
>        
> ---------------------------------
> Ahhh...imagining that irresistible "new car" smell?
>  Check outnew cars at Yahoo! Autos.
> 
> 
> 


-- 
Join FSF as an Associate Member at:
<URL:http://member.fsf.org/join?referrer=3014>