[bksvol-discuss] Re: storage space for Bookshare

  • From: Rogerbailey81@xxxxxxx
  • To: bksvol-discuss@xxxxxxxxxxxxx
  • Date: Sun, 6 Sep 2009 19:59:49 EDT

As a matter of fact, I used to be able to get a list of every book in the 
collection. Before the new site went live I could go to the advanced search 
and leave all fields blank and adjust everything else to return the broadest 
results possible . I would then get a list of every title in the collection. 
At least I suppose it was. At the top it would say 1 to 100 of... and then 
there would be a number somewhere around 50000. I used to do that now and 
then just to satisfy my curiosity about the exact number of books in the 
collection at any given time. Since the new site went live it has been required 
that something be typed into those fields for title or author or key word. 
That has to limit the search in some way that even though it may return a 
large number of results it is no longer the entire collection. 

                                                                          
"The end may justify the means as long as there is something that justifies 
the end. 
" Leon Trotsky     

                 The Militant: http://www.themilitant.com Pathfinder Press: 
http://www.pathfinderpress.com
Granma International: http://granma.cu/ingles/index.html
                 _

table with 2 columns and 6 rows
Subj: 
[bksvol-discuss] Re: storage space for Bookshare   
Date: 
9/6/2009 4:35:22 PM Eastern Daylight Time  
From: 
dbeaver888@xxxxxxxxx  
Reply-to: 
bksvol-discuss@xxxxxxxxxxxxx  
To: 
bksvol-discuss@xxxxxxxxxxxxx  
Sent from the Internet 
(Details) 
table end

Hi all, 

As an old database person let me drop my 2 cents in here. 

The location of the servers makes extremely little difference at all.  
Pretty much the only difference location makes is that the further away from a 
given
user the servers are then responses from the server will seem a tiny bit 
slower. 

Extremely generic searches where the results can be very large lists stress 
database engines due to the amount of resources it takes for the engine to 
process
them.  If Evan, E and I were to request a list that would return most of 
the books in the collection then whomever was first would get first crack at 
those
resources.  The others would have to wait longer and anyone else using the 
sight requesting info via the database would also have to wait until those 
long
requests completed. 

Now, with computers which can do several things at the same time this is 
not as much of a problem.  However, even with multitasking systems the 
database
is still limited on how much it can do in a given amount of time.   

Another issue here is the database being used.  I have no idea what 
database tool Bookshare is using but some of the databases are much slower than 
others
and process data requests in a much more single threaded manner than 
others. 

I believe 

that it is not really our business which servers and databases Bookshare is 
using.  Bookshare has the responsibility to provide the services they 
promise
and to do it as efficiently as they can.  They have responsibility to 
manage the funding they have and to spend it on the best they can within the 
limits
they have.  In order to do this they have to balance funding against costs 
of server space and resources, responsiveness of the web site, service to 
how
many users, etc. 

Personally, I believe they do an admirable job delivering exactly what they 
promise.  They have not promised to deliver entire lists of authors or 
books
either for that matter. 

I might suggest here a way they could deliver the entire list of authors 
and books.  This method would cost Bookshare little in way of space, server 
processing
or database resources. 

Bookshare could develop a process that would execute nightly at midnight 
perhaps.  This process would execute a request to list all the authors or a 
request
to list all books in the collection.  These lists could be written to a 
given place on the server.  It would only be as fresh as once every 24 hours. 

They could then develop a web page where we could go to and download these 
lists.  They could be plain text or .csv or .xls files.  Even if all of us 
downloaded
these files it would put little stress on the servers.  Once we download 
them we could use other tools to search the lists on our own systems. 

As I indicated before this is simply my attempt at explaining things so 
everyone can understand.  I suggest the possible solution only as a 
suggestion. 
Bookshare engineering may come up with their own method of meeting this 
need. 

I sincerely hope this helps someone/everyone understand how these things 
work. 

Thanks.

Other related posts: