02 March 2005

PubSub changed their URI codes: Mud's Tests is pleased to report that PubSub's Bob Wyman is proven to be unreliable

News Flash: PubSub changed their URI coding, removed characters, but the URIs still don't reliably work, as of 2 Mar 2005.

  • PubSub subscription feeds generated five days in advance of a test in Feb 2005, 2 months after the problem was identified in December 2004, continue to fail to correctly identify target strings

  • PubSub subscription results report only ping-updates, not actual content. PubSub now produces no results or content, and provides only the following message in the search results, with the short comment:

    Latest Updates of sites listed in ___
    PubSub fixed their coding problem

  • Their URIs all match

  • PubSub removed the coding problem causing the bookmark-URI to add extra code

    Information

    PubSub gives you four ways to get the URI for the searches. When you add a search, you are given a bookmark; you can also go to your account and check on either the RSS or Atom feed; or you can go to the box on the homepage and click there.

    A few weeks ago, one of those options was not working. Bob Wyman stated emphatically there as no problem, that there was no delay.

    We asserted otherwise and outlined a plan to determine both the cause of the problem and the scope of the error. It took alot of time to figure out why there was a problem. Here's a list of ideas on what was considered.

    This was based on the following test where we said that it did not appear as though the two URIs matched:
    Observation: The XML-feed URI-URL final codes are not similar. [From January 12, 2005]
    Granted, this was not a new problem. The issue was raised on 19 Dec 2004, yet PubSub continued to assert otherwise.

    Mar 2005: All four URIs have the same number of characters

    Today, we see, going into the PubSub URIs that they've changed something. The old problem PubSub was having was that the bookmark option was adding two-extra characters in the URI code. The fix was to remove these extra characters so that all four generated-URIs matched.

  • What they changed

    In our test, we ran multiple search strings through PubSub. To date, these search strings continue to report no content.

    Let us consider a sample search string that an an incorrect URI. We use Search String Five.

    For Search String Five, look at the extra two characters in each of the bookmark URIs:

    005. A. Copied -- 7d0ad241a.xml
    However, the Account was reporting for both the XML and the Box a different URI, which did not have the additional two characters

    005. B. Saved Account -- c7d0ad24.xml
    The problem was that the four codes didn't match.

    005. C. Error -- Added 1a to XML to create copied
    All that had to be done was get the system to remove the two characters from the system that generated for the URI for the bookmark:

    005. D. Correction -- Remove 1a from Copied to ...
    All of the test strings had this problem. All of the bookmarks had the extra two characters. Yet, despite this problem, management continued to assert there was "no problem".

    The solution was easy

    The solution to adjusting all your old not-working URIs was to remove the extra characters. PubSub is now issuing correct URIs.

    Unfortunately, this doesn't say anything about the results. We inputted a test string, saved it to PubSub on Feb 21st, and sent through the string on the 26th, 5 days later. Nothing hit.

    Review

    It was impossible for anyone to assert that there was "no delay." All four bookmarks didn't produce the same code. This was reported in this blog.

    Yet, despite the clear statement of the problem and that the URIs didn't match, we continued to hear that there was "no problem" and "no delay."

    Please, can anyone on this planet explain how two different URIs for the same search could possibly be presented as being evidence of "no delay"?

    What do we know

  • Management cannot avoid the implications of this blog: That valid questions were raised, but ignored;

  • The original discussion about lag time correctly reported a time-related-lag time;

  • There was no basis for management to assert there was no lagtime in the following comment on Jeremy Zawodny's blog;

  • We have yet to hear a credible explanation from PubSub management why the following error report was ignored, dismissed, and the opposite was asserted without foundation;

  • There is a reasonable basis to question whether management is listening to valid ideas and inputs;

  • When PubSub management asserts that an XML-related solution could easily be implemented, the public needs to consider that there will continue to be a notification problem with those not connected to the internet

  • How many other error reports were ignored?

  • What else should we be surprised about?

    Current status

    PubSub has fixed their coding problem. But the fix doesn't resolve the searching-reporting problem.

    Now all four PubSub-generated codes do not correctly work. Before the fix, they had just one. An increase of 4-times, or 500% growth. Wow, are they going to put that on their website?

    BTW: We still have no results for the original search strings entered during the original test. Two months, 60 days, and nothing.

    I remember reading somewhere about a company in the year 2000 that the CEO of the merged company used on another company, called the Geometric Group. Was there not one used on PubSub prior to the merger announcement in November-December?

    Questions for PubSub

  • Why was the coding fixed?

  • What was causing the extra two characters to be added?

  • Why did management assert there was no delay?

  • What was the basis to assert there was "no problem"?

  • If there was "no problem" and "no delay" why the change in the code?

    Answer

    Mud's Tests tested all four options. Someone at PubSub, or the outside QA team did not do this check.

    There was a delay. Bob Wyman asserted otherwise. We no know that when Bob Wyman asserts something, it is best to check.

    Mud's Tests was right. Bob Wyman was wrong.


  • Monitor 110 and PubSub Integration: Are the problems known and disclosed?

    Do Stewart and Parker know about this?

    Have you seen this number on your screen?

    Who might want to know this? CBS Market Watch's Frank Bernako, who covered the PubSub transfusion with the Geometric Group creating Monitor110.

    How "real time" can Monitor110 be when management is chasing bugs in the original platform?

    Has anyone talked to Jeff about this?]

    Take a look at the 2000 Edition of Business Week, reprinted here for educational and discussion purposes only:



    Copyright 2000 Business Wire, Inc. Business Wire
    March 2, 2000, Thursday
    DISTRIBUTION: Business Editors
    LENGTH: 1308 words

    HEADLINE: Independent Testing Confirms Mimeo.com Exact Print Software-SM Gives Users Unsurpassed Reliability; OnDemand Digital Printing and Publishing Conference

    DATELINE: NEW YORK, March 2, 2000

    - - - Snip - - - -

    Mimeo.com contracted QualityLogic, an independent quality assurance and testing company, to put the technology through a thorough testing. According to Randy Kopel of QualityLogic, "Mimeo.com has proven itself a sound and dependable solution for document printing needs. We ran hundreds of files from dozens of applications over a three-day period through the Mimeo.com technology, and found no application errors. We were amazed at the flawless performance."

    - - - Snip - - - -

    "Our best source of ideas is our customers. We learn through their use of our service that Mimeo.com is time-saving and convenient for business professionals in a number of different markets," said Mimeo.com VP Production, Scott Klemm.

    - - - Snip - - - -

    About QualityLogic, "The Quality Assurance Company"

    Formed in September 1999 from the merger of Genoa Technology, Inc.,a world leader in the development of test tools for imaging and telephony technologies, and Revision Labs, a pioneer in software testing services, QualityLogic provides complete quality assurance solutions. Our products and services, under the Genoa and Revision Labs names, are used today by hundreds of OEMs in the computer and communications industries. Many of our products have become accepted as the defacto standard for testing.

    URL: http://www.businesswire.com

    LOAD-DATE: March 3, 2000


    Questions

  • Why didn't Scott get with Bob Wyman to make sure that the testing was done?

  • They had an entire year before the November 2004 announcement of Monitor110 to get this right, yet Wyman continued to assert things were going well. Why?

  • What testing was done and what were the results prior to Geometric and PubSub combining?

  • Was Scott given incorrect information about the status of the PubSub platform?

  • Was someone in PubSub asserting things were "just fine," and convinced someone that "we don't need to spend money on that quality testing stuff"?

  • How many investment companies have shares in Monitor110, and have no interest in publicly discussing problems with PubSub or Monitor110 because sales of the Monitor110 product would mean lucrative financial gains in re IPO stock?

    Put another way, we can see that Monitor110 has a close relationship with the investment community. The customers of the Monitor110 product could very well be the same people who are also the ones providing the funding to monitor Hedge Funds.

  • Is it not possible that material problems with PubSub and or Monitor 110 that could affect an investment company's ability to monitor online discussions would be of interest; yet the investment company would still buy a product that may not work because the company owns IPO shares in both PubSub and Monitor110?

  • In other words, why would the public listen to the feedback from the investment company about either PubSub or Monitor 110 when the "customers of Monitor110" may also be shareholders in the IPO position?

    Summation

    Well, it certainly has been an interesting ride since December hasn't it. The problem was originally raised and each time it was ignored.

    Here we are in March 2005 and now things are falling into place. Wyman had a problem with his platform and emphatically asserted one thing, while the truth was the opposite.

    It remains to be seen whether management's responses to customer feedback [see the 2000 Business Week Article about Mimeo's view on Customer Service] should be taken at face value: That they say one thing and do something else; or whether there was a larger issue at work.

    Perhaps the company was negotiating bridge funding or IPO shares with the customers;

    Perhaps the company was asserting that Monitor110 "should work" because of the great track record of PubSub;

    Perhaps management hoped to bury these questions in the flurry of denials.

    What we do know is Wyman said there was no problem, yet the code has changed. Despite the change in URI code, the PubSub platform continues to have a problem.

    I don't know about you, but it sounds like there's a larger problem than a simple credibility and veracity issue with Bob Wyman. It remains to be seen whether the Securities and Exchange Commission is aware of this, or has any cause for concern, inter alia:

  • 10b-5: Has the public knowingly been given materially false and misleading information;

  • Have investment bankers been asked to keep quiet about the problems in exchange for IPO shares;

  • Is it appropriate for an investment banking firm to be both a customer and a potential investor in a product that may not actually work as advertised; and

  • Is it appropriate for any firm to say they "care" about customers, all the while actively ignoring their feedback and dismissing their concerns and product feedback?

    Let's ask Bob. I'll believe the opposite.
  • News Flash: PubSub changed their URI coding, removed characters, but the URIs still don't reliably work, as of 2 Mar 2005.

  • PubSub subscription feeds generated five days in advance of a test in Feb 2005, 2 months after the problem was identified in December 2004, continue to fail to correctly identify target strings

  • PubSub subscription results report only ping-updates, not actual content. PubSub now produces no results or content, and provides only the following message in the search results, with the short comment:

    Latest Updates of sites listed in ___
    PubSub fixed their coding problem

  • Their URIs all match

  • PubSub removed the coding problem causing the bookmark-URI to add extra code

    Information

    PubSub gives you four ways to get the URI for the searches. When you add a search, you are given a bookmark; you can also go to your account and check on either the RSS or Atom feed; or you can go to the box on the homepage and click there.

    A few weeks ago, one of those options was not working. Bob Wyman stated emphatically there as no problem, that there was no delay.

    We asserted otherwise and outlined a plan to determine both the cause of the problem and the scope of the error. It took alot of time to figure out why there was a problem. Here's a list of ideas on what was considered.

    This was based on the following test where we said that it did not appear as though the two URIs matched:
    Observation: The XML-feed URI-URL final codes are not similar. [From January 12, 2005]
    Granted, this was not a new problem. The issue was raised on 19 Dec 2004, yet PubSub continued to assert otherwise.

    Mar 2005: All four URIs have the same number of characters

    Today, we see, going into the PubSub URIs that they've changed something. The old problem PubSub was having was that the bookmark option was adding two-extra characters in the URI code. The fix was to remove these extra characters so that all four generated-URIs matched.

  • What they changed

    In our test, we ran multiple search strings through PubSub. To date, these search strings continue to report no content.

    Let us consider a sample search string that an an incorrect URI. We use Search String Five.

    For Search String Five, look at the extra two characters in each of the bookmark URIs:

    005. A. Copied -- 7d0ad241a.xml
    However, the Account was reporting for both the XML and the Box a different URI, which did not have the additional two characters

    005. B. Saved Account -- c7d0ad24.xml
    The problem was that the four codes didn't match.

    005. C. Error -- Added 1a to XML to create copied
    All that had to be done was get the system to remove the two characters from the system that generated for the URI for the bookmark:

    005. D. Correction -- Remove 1a from Copied to ...
    All of the test strings had this problem. All of the bookmarks had the extra two characters. Yet, despite this problem, management continued to assert there was "no problem".

    The solution was easy

    The solution to adjusting all your old not-working URIs was to remove the extra characters. PubSub is now issuing correct URIs.

    Unfortunately, this doesn't say anything about the results. We inputted a test string, saved it to PubSub on Feb 21st, and sent through the string on the 26th, 5 days later. Nothing hit.

    Review

    It was impossible for anyone to assert that there was "no delay." All four bookmarks didn't produce the same code. This was reported in this blog.

    Yet, despite the clear statement of the problem and that the URIs didn't match, we continued to hear that there was "no problem" and "no delay."

    Please, can anyone on this planet explain how two different URIs for the same search could possibly be presented as being evidence of "no delay"?

    What do we know

  • Management cannot avoid the implications of this blog: That valid questions were raised, but ignored;

  • The original discussion about lag time correctly reported a time-related-lag time;

  • There was no basis for management to assert there was no lagtime in the following comment on Jeremy Zawodny's blog;

  • We have yet to hear a credible explanation from PubSub management why the following error report was ignored, dismissed, and the opposite was asserted without foundation;

  • There is a reasonable basis to question whether management is listening to valid ideas and inputs;

  • When PubSub management asserts that an XML-related solution could easily be implemented, the public needs to consider that there will continue to be a notification problem with those not connected to the internet

  • How many other error reports were ignored?

  • What else should we be surprised about?

    Current status

    PubSub has fixed their coding problem. But the fix doesn't resolve the searching-reporting problem.

    Now all four PubSub-generated codes do not correctly work. Before the fix, they had just one. An increase of 4-times, or 500% growth. Wow, are they going to put that on their website?

    BTW: We still have no results for the original search strings entered during the original test. Two months, 60 days, and nothing.

    I remember reading somewhere about a company in the year 2000 that the CEO of the merged company used on another company, called the Geometric Group. Was there not one used on PubSub prior to the merger announcement in November-December?

    Questions for PubSub

  • Why was the coding fixed?

  • What was causing the extra two characters to be added?

  • Why did management assert there was no delay?

  • What was the basis to assert there was "no problem"?

  • If there was "no problem" and "no delay" why the change in the code?

    Answer

    Mud's Tests tested all four options. Someone at PubSub, or the outside QA team did not do this check.

    There was a delay. Bob Wyman asserted otherwise. We no know that when Bob Wyman asserts something, it is best to check.

    Mud's Tests was right. Bob Wyman was wrong.


  • Monitor 110 and PubSub Integration: Are the problems known and disclosed?

    Do Stewart and Parker know about this?

    Have you seen this number on your screen?

    Who might want to know this? CBS Market Watch's Frank Bernako, who covered the PubSub transfusion with the Geometric Group creating Monitor110.

    How "real time" can Monitor110 be when management is chasing bugs in the original platform?

    Has anyone talked to Jeff about this?]

    Take a look at the 2000 Edition of Business Week, reprinted here for educational and discussion purposes only:



    Copyright 2000 Business Wire, Inc. Business Wire
    March 2, 2000, Thursday
    DISTRIBUTION: Business Editors
    LENGTH: 1308 words

    HEADLINE: Independent Testing Confirms Mimeo.com Exact Print Software-SM Gives Users Unsurpassed Reliability; OnDemand Digital Printing and Publishing Conference

    DATELINE: NEW YORK, March 2, 2000

    - - - Snip - - - -

    Mimeo.com contracted QualityLogic, an independent quality assurance and testing company, to put the technology through a thorough testing. According to Randy Kopel of QualityLogic, "Mimeo.com has proven itself a sound and dependable solution for document printing needs. We ran hundreds of files from dozens of applications over a three-day period through the Mimeo.com technology, and found no application errors. We were amazed at the flawless performance."

    - - - Snip - - - -

    "Our best source of ideas is our customers. We learn through their use of our service that Mimeo.com is time-saving and convenient for business professionals in a number of different markets," said Mimeo.com VP Production, Scott Klemm.

    - - - Snip - - - -

    About QualityLogic, "The Quality Assurance Company"

    Formed in September 1999 from the merger of Genoa Technology, Inc.,a world leader in the development of test tools for imaging and telephony technologies, and Revision Labs, a pioneer in software testing services, QualityLogic provides complete quality assurance solutions. Our products and services, under the Genoa and Revision Labs names, are used today by hundreds of OEMs in the computer and communications industries. Many of our products have become accepted as the defacto standard for testing.

    URL: http://www.businesswire.com

    LOAD-DATE: March 3, 2000


    Questions

  • Why didn't Scott get with Bob Wyman to make sure that the testing was done?

  • They had an entire year before the November 2004 announcement of Monitor110 to get this right, yet Wyman continued to assert things were going well. Why?

  • What testing was done and what were the results prior to Geometric and PubSub combining?

  • Was Scott given incorrect information about the status of the PubSub platform?

  • Was someone in PubSub asserting things were "just fine," and convinced someone that "we don't need to spend money on that quality testing stuff"?

  • How many investment companies have shares in Monitor110, and have no interest in publicly discussing problems with PubSub or Monitor110 because sales of the Monitor110 product would mean lucrative financial gains in re IPO stock?

    Put another way, we can see that Monitor110 has a close relationship with the investment community. The customers of the Monitor110 product could very well be the same people who are also the ones providing the funding to monitor Hedge Funds.

  • Is it not possible that material problems with PubSub and or Monitor 110 that could affect an investment company's ability to monitor online discussions would be of interest; yet the investment company would still buy a product that may not work because the company owns IPO shares in both PubSub and Monitor110?

  • In other words, why would the public listen to the feedback from the investment company about either PubSub or Monitor 110 when the "customers of Monitor110" may also be shareholders in the IPO position?

    Summation

    Well, it certainly has been an interesting ride since December hasn't it. The problem was originally raised and each time it was ignored.

    Here we are in March 2005 and now things are falling into place. Wyman had a problem with his platform and emphatically asserted one thing, while the truth was the opposite.

    It remains to be seen whether management's responses to customer feedback [see the 2000 Business Week Article about Mimeo's view on Customer Service] should be taken at face value: That they say one thing and do something else; or whether there was a larger issue at work.

    Perhaps the company was negotiating bridge funding or IPO shares with the customers;

    Perhaps the company was asserting that Monitor110 "should work" because of the great track record of PubSub;

    Perhaps management hoped to bury these questions in the flurry of denials.

    What we do know is Wyman said there was no problem, yet the code has changed. Despite the change in URI code, the PubSub platform continues to have a problem.

    I don't know about you, but it sounds like there's a larger problem than a simple credibility and veracity issue with Bob Wyman. It remains to be seen whether the Securities and Exchange Commission is aware of this, or has any cause for concern, inter alia:

  • 10b-5: Has the public knowingly been given materially false and misleading information;

  • Have investment bankers been asked to keep quiet about the problems in exchange for IPO shares;

  • Is it appropriate for an investment banking firm to be both a customer and a potential investor in a product that may not actually work as advertised; and

  • Is it appropriate for any firm to say they "care" about customers, all the while actively ignoring their feedback and dismissing their concerns and product feedback?

    Let's ask Bob. I'll believe the opposite.
    " />