Entries:
Comments:
Posts:

Loading User Information from Channel 9

Something went wrong getting user information from Channel 9

Latest Achievement:

Loading User Information from MSDN

Something went wrong getting user information from MSDN

Visual Studio Achievements

Latest Achievement:

Loading Visual Studio Achievements

Something went wrong getting the Visual Studio Achievements

Discussions

figuerres figuerres ???
  • What's with all the complaining about high DPI displays for Windows desktop?

    It's a mess and yes the reasons run deep in how all kinds of apps have been written with developers thinking for the most part that we are still back in the 72dpi bitmap world.

    i hate and love when read of changing font size when they get a high rez screen.

    really the size of a 72 point font is 1 inch. that has zero relation to the ppi of the display!

    but then we create a form based on pixel sizes and set the text based on point sizes in most cases.....    creating a built in failure when that relationship changes.

    also this is one of the things MS was trying to get us away from when they first created WPF and vista.

    do the display based on drawing with math not on counting pixels!

    make form sizes base don inches or centimeters or points not on pixels

    same thing for buttons and other UI elements.

    only use a bit map to show a picture of something, not to create a UI element like a button.

    we need to do this kind of thing all over the code base, bottom to top.

    if the display is based on the math then you can scale it to fit the display area and it works.

    in fact i have used a wpf control to scale forms and shown how it can work very well. in a wpf window with wpf markup inside the container.

  • Internet or a web of networks

    , ScanIAm wrote

    It's kind of obnoxious that the billing doesn't reflect the actual 'cost' in any meaningful way.  The difference in cost of sending 100 bytes vs 1000 bytes isn't an order of 10 no matter what kinds of costs you try to include, and yet providers of all forms try to convince us that there is a correlation.  I get that it's hard to keep everybody happy when it comes to billing, but every 'reason' for billing to be the way it is can be summed up by 'because we can'.

    well there are some sunk hardware and network costs to scale up the transfer but that is a kind of "one time cost" that should be allocated across all of the customers and the time it will be in service.

    but also of note is the way fiber has tended to get more data thru the same fiber every few years by better encoding of the light signal.

    that has kept the backbone providers from having to dig up more trenches and adding more fiber to grow to meet demand as much as if they were using copper to connect.

    there are a lot of details in how it all works and yes by and large it's all about how much money you can get w/o spending more money.

  • Internet or a web of networks

    , davewill wrote

    *snip*

    From what they have stated that does seem to be the norm still today.  What doesn't seem to be the norm though (at least in this case it sounds like it isn't) is when the network traffic is out of balance the affected party should bill the provider on the other side of the edge.  That provider then in turn bills up their stack.  Apparently that isn't the norm today I gather.

    back around 1996-2000 when I dealt with more of this stuff there were a number of billing formulas used to bill a customer, generally based around finding the "normal peak" use for a month.  but what I dealt with was ISP to user or ISP to web site; not Backbone peering.

  • Internet or a web of networks

    , davewill wrote

    What is perplexing me is why the market economics don't flow back up the stack.

    customer -> last-mile ISP -> backbone ISP -> Netflix

    Unless Verizon wants to be both a last mile and a backbone network.

    I suspect that part of the problem is that Verizon has this business under the "FiOS" brand to sell cable tv to subs and Netflix may be seen as taking from that revenue stream.

    not to be "captain obvious" but that is part of the thing...

    also adding complexity is that they are not one company, by the FCC rules they have multiple corps. so there are at least :

    Verizon

    Verizon Wireless

    Verizon Internet

    Verizon FiOS

    also if the internet peering billing is done the way they used to do Telco peering it's based on even sharing of traffic and getting the other peer to pay more if they send more.

    at one time Telco billing used to be such that if you "terminated" a call from another Telco then that Telco paid you for finishing the call.

    but I bet that Netflix does not want to do that as it would cost them way to much.

     

  • Extracting the version number from a cab file using C# .net Compact Framework

    , Varunkrishna wrote

    @cheong:and @ figuerres

    First and foremost I would like to thank you both for answering my questions.

    So I have now a xml file with the nodes something like this

    1
    2
    3
    <app>
    <versionNumber>0.0.0.0</versionNumber>
    </app>

    so with this type of thing can/how (do) I check whether the installed CE app's version is same as the one on the server ?

     

    Thanks

    Varun Krishna. P

    I do not have the code in front of me to check but what I did was a .Net call to get the running assembly's version.

    this link has the info:

    http://www.stevenalowe.com/how-to-get-the-application-version-in-net/

     

     

  • Opening txt files in Pocket Pc

    , Varunkrishna wrote

    @figuerres:

    my app stored error and log info in a format In which format do you store, log.txt or something like this log file's content in sql DB ? sorry  I could not get you.

     

    Thanks

    Varun Krishna. P

    if you re-read my post I said that I had a web service and uploaded the log data to that service.

    on the server the data went into a sql database.

    I had around 80 hand held devices deployed remote so local log files were not much use.

    what I did was used a library that added binary serialization to .net CF and I used that. when uploading the data it was in a SOAP / XML format for an ASMX web service.

     

  • Opening txt files in Pocket Pc

    what I did:

    create a web service that stores logs on a sql db.

    my app stored error and log info in a format I could later upload via the web service.

    then I could check the db for logs. that way if a user has a problem they can get in range of WiFi or 3G and the app will upload the log files for the user automatically.

    and you can collect data from multiple users to decide what errors to fix.

    just have to make sure that the core "Get online and upload errors" works really good.

     

  • Extracting the version number from a cab file using C# .net Compact Framework

    , Varunkrishna wrote

    Hi all

    I have a pocket pc application , I am going to upload the cabinet (.cab) file on to a local server, I would like to check/extract the version number from the cab file. Is that possible to extract the cab file's version using c#.netCF ?

     

     

    Thanks

    Varun Krishna. P

    when I maintained some apps for ce I named the files with the version number.

    I also had an app that would download the current version and update the main app.

    the device would run the updater app - it would check and download if it needed an update and then run the up-to-date main app.

    I used a zip lib and stored updates in a zip file so that updates to the .net app were simple xcopy updates.

  • Can VS create a visual graph of typical high performance code?

    OK I think it's funny how you keep writing "High Perfomance" but you are not wanting to run the code to get that data.

    it might be, it might not be. you do not really know unless you test it.

    but that aside ....

  • Can VS create a visual graph of typical high performance code?

    sounds like you want to make a graph based on two things:

    1) a class diagram

    2) profile the code

     

    then use the profile data to annotate the class diagram.

    IMHO your question should be re-phrased more like this

    how to create a graph that identifies frequently called methods in a class or library.

    the "high performance" has nothing to do with finding the frequently used methods in a class.

    VS ultimate has what I think you want, it will do this.

    also any graph like this will be different for each program calling that library / class and may give false results unless you profile a number of calling programs over a number of runs each.

    like the expression "YMMV"